Conference 7.286::digital

    OK I tried it. I assume, tongue half in cheek, that the purpose
    is to drive people back to the paper phone books.

    I found a topic, 64, in IAMOK::ELF with the "most commonly asked
    questions" for ELF V2. The reasons for replacing V1 seemed to be
    quite valid. They didn't really address the "benefits" of the
    user interface itself other than to say that most people know how
    to use VTX. A contention I would be willing to dispute. I'm still
    reading topic 59 there which discusses usability.

    		Alfred
    PS: KP7 etc to add the ELF V2 conference IAMOK::ELF to your notebook.

Re .0

  I disagee. I prefer the VTX version to the V1 version. At least with
  VTX you don't have to keep retrying to get an elf server.

  A couple of points though:

  Elf V1 had the very useful facility of letting you enter your own
  message. Elf V2 only lets you enter short replies to specific fields
  such as your location, telephone number etc. Almost without exception,
  anybody who has never come across my surname before will pronounce it
  completely wrong. So, I put a line in the Elf V1 message field telling
  people who try to find my phone number there how to pronounce my name
  BEFORE they call me up. It's useful having an 'open' field like this
  where people can put in what they like, rather than all the pigeon
  holes we've got now. Oh well, maybe I'll just need to become famous :-)

  Another thing. Our Elf V1 server tells me peoples badge numbers. Elf
  V2 doesn't. Has this suddenly become classified information?

  There's a notes conference about ELF V2 on IAMOK::ELF which may be
  more suitable for this discussion.

	Craig. 

    RE: .2 I found a command file called ELF_RETRY that keeps trying
    ELF servers until it gets one. I haven't had any complaints using
    ELF V1 since.
    
    		Alfred

tell us how to pronounce it!

Bruce

Not to mention the fact that in <SOME> parts of the country
(i.e. Western Area), DIS hasn't set up VTX properly, so 
VTX <keyword> doesn't work out here.  I'm now going to
have to wend my way through n levels of VTX menus before
I get the joy of disappearing further into the quagmire?????
I cant wait!

Kevin

    There is something available from the TOOLSHED (I believe is called
    EL2) that will allow you to use V2 of ELF on a single command line
    without entering VTX directly.
    
    The few things I find annoying about the "new improved" ELF is:
    
    1)  If I try to find ROBERT NEWMAN (my "formal" name) it will not find
    me but will find other ROBERT NEWMAN's within Digital.  If I ask for
    BOB NEWMAN it will find me.  I know this is how I am in the phonebook
    and other places but I find it annoying.
    
    2)  The old ELF seemed to give you a "best guess" to someone if you
    misspelled their name.  The "new" ELF does not do this.
    
    Maybe I am just using it wrong but I like the old ELF better and with
    ELF_RETRY mentioned in an earlier note (also available from the
    TOOLSHED) I have no problems with V1

    I'm willing to give it a try, based on the fact that VTX should be more
    efficient than the old method, at least allowing me to even GET to ELF
    when I need/want to.
    
    However, in updating my own entry, I find the user interface about as
    friendly as a rabid dog.  For example, on the "user updateable info"
    screen, the FIRST thing I did was to add "JOHN" as a searchable given
    name.  (Why do people insist on putting an "H" in "JON"?)  This was in
    my nickname field in ELF V1; wonder why it wasn't carried over to V2?
    
    Why does it ask for "organizational unit" when it doesn't allow any
    special characters like "/" or "-"??  "SWS/E" I can read and
    understand, but there ain't enough room to spell it out, even if I was
    so inclined, which I'm not!  And when it claims you've made an error,
    it puts the cursor back to the FIRST enterable field on the form
    instead of the field with the error.  Who was it that made mention of
    1975 technology being used...
    
    Finally, TWO DAYS for this more efficient system to get the master file
    updated with my changes?  Didn't the old one do it overnight?
    
    sigh.....
    Jon

       
       
       	Moderators, please let us know definitivly if we can continue
        discussing this here.. 
       
       	I have to agree with just about everyone.. ELF V2 is such a
       backasswards interface to something that was so simple.. The
       BIGGEST blunder was in removing the ability to put a personal
       message. How am I going to tell the world that I've been
       transfered to the DEC office in Thule, Greenland or that I'm
       out of the office on special assignment for KO? Seriously, 
       ELF V2 was never designed with humans in mind.. (AND DEFINATELY
       not with engineering types in mind!!)
       
       							mike

    For now, the discussion can continue here, but personally, I think
    it would be more productive in IAMOK::ELF - that is, if people want
    the hope of accomplishing something instead of just venting.
    
    	Steve (who vented on this topic a long time ago...)

    I thought the new elf was so bad, I've stopped using it.  Maybe that 
    was the design goal.

All of the answers in the ELF notesfile seem to be either:

	1. Read the manual
	2. Enter a more complicated string ("/G=foo/S=bar/L=xyo")
	3. It's DDS' fault, not ELF
	4. V1 was so bad we had to make it worse

Jon Arnold will find that adding "JOHN" to his list of search given names
won't help much.  FIND JOHN ARNOLD will only return the "real" John Arnold",
to find _him_ you'd have to type FIND /G=JOHN/S=ARNOLD but that won't work
because DDS retrieves ALL THE JOHNs IN THE COMPANY, and ALL THE ARNOLDs, and
then computes the intersection of the two sets.

FIND BOB SMITH  and   FIND ROBERT SMITH    will always find two separate sets
of people, and there is no way to fix this for very common last names like
Smith.

Some other things to note:

	1. You can't change your MS/Loc (called Intrnl Mail Addr)
	   yourself, only your PSA can change it.
	2. Same thing with your phone number.
	3. Additional fields (like additional phone numbers, which you
	   _can_ put it) are *NOT* displayed without a "/FULL" qualifier
	   on the FIND command.

And finally, unlike a new version of a piece of software you don't like as
well as the old one, YOU CAN'T IGNORE this change.  After 9 November, the
old ELF will refuse to accept changes, and after some time in December, all
the old ELF servers will be shut down.

/john

    Can someone confirm what is described in .6 2), which is the fact
    that VTX elf is unable to find a "best guess" ?
    
    I cannot believe it !
        
    In order to know the EXACT spelling of a name, I will be forced
    to look into a book, which will automatically gives me the DTN and
    ALL-IN-1 account, just what I need. 
    Therefore, what the remaining/new advantages ?
    
    Thierry

ELF V2, "because it is based on Digital's DDS (Distributed Directory Service)
product" can only find exact matches.

You _can_ use a _trailing_ "*" to cause a wildcard match.  When you do wildcard
matches, the names are _not_ returned in a sorted order.

One of the REALLY bad things is that if you know that someone's first initial
and last name, you may or may not be able to find them.  FIND J FOOBAR will
work in some cases, usually times out.

The answer in the ELF conference is "provide as many characters as possible,
followed by an asterisk."

/john

              
    There are two reasons for the change, I think:
                
    The first reason for the change (I used to work with the guy who
    had to SUPPORT the V1 elf hack) was that ELF V1 was a maintenance
    nightmare. 
    
    The second is that DIGITAL wanted to be able to sell elf-like
    functionality to our customers and V1 was unsupportable, thus
    unsellable. And of course, we use what we sell, right?

        Re: .all

        I have no idea why you guys are complaining about the new ELF.
        It just finished it's field test and no significant problems
        were found.  From the field test report:

        		Priority 1 - 0
        		Priority 2 - 0
        		Priority 3 - 0
        		Priority 4 - 0
        		Priority 5 - 5

        Some problems were found for priority 1-4, but they were all
        fixed during field test.  None of the complaints mentioned in
        the previous replies were listed as problems, so they must not
        be problems.

        

        I wish I could put a smiley face on the preceding text, but it
        seems to be the true and accurate view of the world as seen by
        the ELF developers.

        I know that ELF V1 has problems.  The new version of ELF has
        been `on its way', and I've been looking forward to it, for a
        long time:

         o  Note 111.4 in the old ELF conference (MSTEAK::ELF) says
            `While we wait the year or two for DDS to become useful, 
            ELF will still be out there plugging away.'  This note 
            was written 2-NOV-1984.

         o  Note 254.1 from 18-MAY-1987 in the old conference talks
            about a version of ELF under development that will use VTX. 
            That's right, it took over 2.5 years to develop and inflict
            (or, as they say, `implement') this hideous user interface.

        But now that ELF V2 has arrived, and I wish we could kill it 
        and start over again.

        Note 57.2 in the new ELF conference (IAMOK::ELF as previously
        mentioned) lists Enterprise Network Applications Services
        Management Group representatives for various parts of the
        corporation.  They didn't develop ELF, but they supervise it 
        and made the final decision to `implement' the new version.

        					B.J.

    Gee, B.J., I *think* there's a differnce between no 'bugs' and no
    problems.  I used the old ELF [did you ever try that one?], and
    I just tried the new one.  I may be supportable now in more ways
    than one.  Who from a user standpoint [having used the old one]
    would want to use the new one?  Its kinda like haveing to use
    VMS after having the privilage of using TOP-20s ?
    just my opinion.
    ed
    

>    What possible justification is there for killing off the barely
>    adequate "ELF V1" as of December 18 and replacing it with the
>    grossly inadequate "VTX ELF"?

    'cause a whole bunch of people spent years developing it.  You wouldn't
    want them to think it was all a waste of time simply because it's
    no good, would you?

    I try to make it a rule never to judge an individual contributor unless
    I've walked a mile in his management situation, but really,
    two-and-a-half years to develop V2 because V1 was a "maintenance
    nightmare"? 

    ELF just isn't that big a deal as far as a piece of software goes.  If
    it's a maintenance nightmare, scrap it and reimplement it.  My group
    just did that with a much more complex piece of software.  All over the
    company, single engineers are heroically maintaining, often
    reimplementing, real hairy bears of software, task builders, compilers,
    linkers, various pieces of VAX Document (my product).  The ones I know
    would consider ELF to be an interesting hack, not a program at all.  In
    fact, until this came up, I thought that's what ELF was.

    If the servers were a big problem, add more.  Is it harder to be an ELF
    server than a notes server?  There are plenty of notes servers.

    I'm not surprised that VTX ELF is based something called DDS --
    using it's about as enjoyable as going to the dentist.

    I am not sure how many of you were around few years ago.  I am talking
    about the early eighties.  We were going to build the best personal
    computer based on venerable RSX and we were going to make it very easy
    to use.

    From what I learn afterwards (I joined the group little later), the
    engineers working on the project considered themselves the God and
    would not budge even though users were complaining.  People hated those
    damn slow menus.  They had to stare at "Returning to Main Menu" for a
    long long time.  Anyone care to remember the heated debates about <DO>
    vs <RETURN>??

    Then came KNotes, Notes-11, Vnotes and the product VAXnotes.  We now
    expect any digital developed product to have the quality and the ease
    of usability that VAXnotes has.  Unfortunately, not all the developing
    groups have the humility to let their product withstand the scrutiny of
    ENet community as far as their development effort is concerned.

    - Vikas

    Reading this topic and the IAMOK::ELF notes file is an introduction to
    what has been called "The Baby Duck Syndrome". The first animate object
    that a baby duck sees when it hatches is mother. Similarly the first
    word processor, mail system or, in this case, directory service which a
    user learns is the best, the world's finest and only a person of very
    low intelligence could fail to appreciate it's magnificence. The fact
    that it (WP, EM or DIR) could only be used by something with two webbed
    feet and a beak is irrelevant.
    
    It appears to be a fact of computing life that many end users are
    totally devoted to the interface that they first learn. They are
    extremely resistant to any change, no matter how logical or illogical,
    beneficial or harmful. All change is bad and caused by those of less
    perception, intelligence, creativity and moral standing than that of
    the affected user. This "fact" leads to countless mindless "wars". Does
    anybody remember the "keyboard wars" around the LK201?
    
    No doubt the developers of VTX, DDS and ELF Version 2.0 could have done 
    some things better. It wouldn't have mattered. The Baby Ducks' feathers
    will still be ruffled. Each one of us could come up with what we, but
    noone else, would consider a better design. Just try and implement it
    and I'll watch you being fed to the ducks.
    
    There is a not very well appreciated problem in designing end user
    applications. It's quite straightforward to develop a user friendly
    easy to learn flexible distributed application. It's extremely
    difficult to scale that application so that it serves a quarter of a
    million users as well as it serves one user.
    
    Those of you who used the original Corporate EMS system (MUMPS based)
    know that it had some very desirable features such as address
    validation on entry, shared distribution lists, etc:. EMS choked to
    death as the user population approached ten thousand. Similarly ELF V1
    was introduced to fill the void for directory service but it is now
    choking to death as the activity level rises.
    
    Have you never thought why you only have one choice (used to be two but
    rotary dials are no longer made) of user interface for your telephone?

Re: .-1,

I'm sorry, but you're way off base.
If I minded learning new interfaces, new products, and new ways of
doing things, I might as well retire, because this is no sort of industry
for that kind of attitude.

The issue here is that the "new, improved" version has demonstrably LESS
FUNCTIONALITY than the previous version.  I have no problem with learning
new UI's, but new is not necessarily better.  In this case, it is much
more cumbersome for most of us.  Nevertheless, that is not the main issue.
If the product is implemented cleanly (like the notes server is) then someone
will invent a front-end which by-passes VTX (like Enotes bypasses TPU).

How would our customers react if we significantly reduced the DCL syntax
in the next release of VMS?

The bottom line is if you take functionality away without giving
significant benefits, you will have a very hard time convincing anyone
that it is an improvement.

Kevin

.21 is total bologna.  We're not talking about preference, we're talking about
nearly total loss of functionality.

ELF V1 finds employees.

ELF V2 doesn't.

It finds less employees (doesn't deal well with given names vs. nicknames for
people with the same name as someone else in the company), times out on totally
reasonable searches (first initial plus last name), doesn't allow data to be
updated in a timely fashion (location and phone number), doesn't display the
additional user-provided location and phone number unless the user knows the
magic "/FULL" incantation, and doesn't provide a place for a status message.

In short, unlike the telephone's keypad dial which does EVERYTHING the rotary
dial did, but faster, ELF V2 falls drastically short of the capabilities in ELF
V1.

/john

In the management overview of ELF V2 (IAMOK::ELF, note 62.1), the author writes:

> While ELF V2 offers many functional, accessibility, and network 
> performance  enhancements, it also has some acknowledged drawbacks: ...
> these can be addressed through user training and awareness.

As I read this, ELF management believes that, if the system doesn't work for
the people, fix the people.

We (Digital) built this company in the idea that computers were a tool
to be used by people, not the other way around.  All of us who have been
here for a few years remember the cartoon series of the swing hanging
from a tree: how it became progressively more complicated as it passed
from marketing to engineering to production to installation.  The last
panel showed a rope and a tire: "what the customer wanted."

From a quick reading of IAMOK::ELF, it seems that it was put together using
tools that were c hosen because we sold them (or want to sell them), and not
because they were the best tool for the task.  In 1973, we sold a PDP-11/20
to the Copenhagen Telephone Company to use for directory access.  Surely,
16 years later, we could do at least as well with our current systems.

If we can't, we need -- as a company -- to understand why, and what this
means for our survival.

Martin.

Re .22:

>If the product is implemented cleanly (like the notes server is) then someone
>will invent a front-end which by-passes VTX (like Enotes bypasses TPU).

If the history of ELF V1 and VAX Notes are any indication of the way things will
go, the protocol will be jealously guarded by the implementors and hackers will
have to reverse engineer it to make the service available on additional
platforms.
				/AHM

    Just be thankful that this project was not implemented as a fixed
    price project for a real customer.  Digital would have lost its
    shirt...
    
    Jon

    John Covert hit the nail on the head: ELF V2 doesn't find employees,
    (unless you know everything about them of course, but in that case
    you'll know their phone no., location and Enet address).
    
    I'd settle simply for SOUNDEX being implemented.
    
    - Andy ��� Leslie
    

       
       
       	I'd settle for management group who signed this catastrophe off
       	to go somewhere's else. Like maybe IBM.
       
       	It never ceases to amaze me that we have people in this company
       that insist on doing things the hard way, no, correct that, go out
       of their way to make things difficult or unworkable for the rest of
       us that need to get something done. Who are these people???
       
       	Hey, maybe it's time to form a revolt. I'm sure that a few of us
       in Engineering could hack up something that works. I can get the
       party together and we can do it some Saturday. (and documented
       by Sunday)
       
       							mike

    Sadly you'd need the co-operation of Personnel for access to the
    records.
    
    - Andy ��� Leslie
    

    Not if you design it so that users can populate it themselves :^)
    
    /Harley

Let's not dwell on the VTX user interface issue; there already is a command
line interface available (ELF_FIND; it's in the toolshed).

It provides access at the DCL level to ELF V2 (for FIND, but not for changes).

But it, too, is totally inadequate in terms of finding anyone, because the
underlying directory service (DDS), is horribly less able to support most of
the kinds of searches people want when looking things up in a directory
than ELF V1 was.

Of course, DDS is fully CCITT X.500 compliant, so it met its goals and its
implementors, in the U.K., all got their raises this year.

Doing anything about ELF V2 is probably too late.  Implementation is two weeks
away.  As of that date, ELF V1 data will no longer be updated (neither by
personnel feeds nor by ELF V1 MODIFY or REGISTER commands).

/john

P.S.: designing something for self-population would be a waste.  The majority
of employees wouldn't bother.

    RE .28
    
    Mike, count me in.  The way I see it (assuming this group could
    get the data leads from personnel), you're right; do it Saturday,
    test & document Sunday, and still have time for dinner & movies
    Sunday evening.
    
    Jon

n
Hmm, does "find * *" return all employees (in a finite amount of time)?

Martin.
ps: this can, of course, be done recursively.  Amazing how computers
can simplify our efforts.

    
    I was looking for a place to gripe about ELF V2.  
    
    Any user interface that forces you to use "PF1" "enter" when you
    should use a <return> is broken.  It reminds me of the saying...
    Make an interface so 'simple' that idiots can use it and only idiots
    will use it.

    Any phone directory program that can't do quick lookups of name
    that is a close match is a real joke.  If I call 212-555-1212 and ask 
    for any of the 8,000,000+ people in 212 area code they get the answer 
    back to you in seconds.  I often have the spelling messed up too!
    
    I have an idea..  We (DEC) buy a copy of the phone company directory 
    software  (I saw it at DECworld on demo. It ran on a MicroVAX).  
    We rig up the DTN to have 411 connect us to a room full of operators 
    that lookup phone numbers and nodenames for us. This can easily scale 
    up to a company of 8,000,000 people by the example of the NYC
    directory assistance.

    This should work great and provide a well known human interface.

    I would gladly use the phone book if it had node::username listed
    for the folks in the book.  Folks that aren't in engineering have
    been brainwashed into thinking that ALL-IN-ONE mail is easier and
    better than VAX mail.  

    The VTX program should have just been a menu driven interface to
    the old style ELF.  You could have the server use the top few lines
    of the message to hold all the new information.
        
        

    Let me add my two cent's. I see only one problem with the new ELF
    That is I can not put a message in as I could with the old one.
    
    The reason I will miss this is I work 3rd shift as a resident,
    and in the old ELF I was able to put a message that said that
    and how to reach me on site.
    
    I have also used this message space to say that I was at training
    or on vacation, that way if someone was looking for me they know
    that I always keep my status updated in ELF.(I.E. ELF>Find Ken Romanik)
    
    But I will not be able to do this with the new ELF. Or should I
    say that if I can I have not seen any info or help commands yes
    
    K.J.R.

	"They" statred field testing ELF V2 in Marlboro about two months
	ago.  I was one of the ones that complained that is was literally a
	step backwards in functionality.  It fell on deaf ears.  They were
	more interested in what was broken than opinions on how things
	worked.  (I suppose I could have argued that, given the way it
	works, it is "broken"--totally broken ... oh well.)

        The gist of the reply was that most of the company is no longer
        made up of "technoweenies" (to borrow a term from a recent
        "Information Week" article) and that the market being addressed is
        people who don't really understand, nor think they should
        understand computers (the "I build computers, I don't use them"--to
        steal a quote from our president--mentality).  Also, there is a
        legitimate issue of the availability of ELF servers versus VTX
        servers, the former being more and more frequently unavailable when
	you need them.  From my reading there is apparently a command proc
	to solve that problem however awareness of such a proc seems low.

        I'm not going to argue the implications of the reply as it leads to
        a bottomless rathole, however, I agree with the sentiment of past
        replies: 

	1) it will force increased use of the paper phone book
	2) someone will figure out a better interface (anyone got any
	ideas?)

For what its worth,
 jim_who_considers_himself_a_"technoweenie"_despite_an_apparent_decline_
 in_its_popularity
	

But the old ELF was much more easily usable by non-technoweenies than this
new abomination!

The old ELF usually found someone if you just entered ELF FIND first last.

The new ELF doesn't.  You have to know magic incantations with slashes and
equal signs and all sorts of things only technoweenies understand!

/john

        Re: .31 (John R. Covert)

> Doing anything about ELF V2 is probably too late.  Implementation is two weeks
> away.  As of that date, ELF V1 data will no longer be updated (neither by
> personnel feeds nor by ELF V1 MODIFY or REGISTER commands).

        When I talked to the MEM representative on the Enterprise
        Network Applications Services Management Group (the committee I
        mentioned in .15), I was told that some of the votes to go ahead
        with implementation were conditional--and it didn't look like
        all the the conditions would be met.

        I'm not optimistic (the conditions will probably be waived to
        get ELF V2 implemented), but it might be possible to do
        something about V2.

        					B.J.

    Point of personal privilege:  I am an English major and writer by
    trade.  I helped start the DCL Clearinghouse and I've been involved in
    user-interface design for all the products I've ever worked on.  In the
    course of my work I have learned MCR, DCL, MS/DOS, the Symbolics
    Command Processor, the Pro350 thing, and A-to-Z, and three versions of
    EDT, EVE, Zmacs, XLSE, plus three or four word processors, not to
    mention a bunch of other software, including ELF V1 and ELF V2, and I
    can tell you that ELF V2 is the most unresponsive, unrewarding,
    spiky-looking, nonfunctional piece of 1965 UI programming I've ever
    seen.  And embedding it in VTX just compounds the crime.
    

	This is from the ELF conference. It explains the "why" of the
	new implementation. It also list the names of the committee that
	sponsors the new ELF. (I've added VMS mail addresses to the list
	as they were missing from the original note.) I assume you could
    	contact them directly if you have major concerns.

			Alfred
================================================================================
Note 62.1                >Management overview: the whys                   1 of 1
IAMOK::DEVIVO "Paul DeVivo @VRO, DIS/DT, 273-5166"   71 lines  12-OCT-1989 10:16
--------------------------------------------------------------------------------
                  ELF VERSION 2 - A MANAGEMENT OVERVIEW


This paper is designed to answer anticipated questions about ELF Version 
2 from a management point of view: why it is being introduced, and why it 
differs from ELF V1.

As most ELF users are aware, ELF Version 1, on the network since 1983, is 
not providing an adequate level of service to employees.  STG/Digital 
Telecommunications has developed ELF V2 to address these deficiencies.  
The project is sponsored by the Enterprise Network Applications Services 
Management Group (ENASMG) to which STG, MEM, and each geography has 
representation:

		Europe: Gabriel Barta @GEO GVA01::BARTA
		GIA: George Faucher @AKO   AKOV11::FAUCHER
		MEM: Lou Jacques @MRO      VLNVAX::LJACQUES
		NNE: Robert Yost @VRO      
		STG: John Beirne @VRO      IAMOK::BEIRNE
		US: Bob Bradbury @PKO      FDCV06::BRADBURY

The objective since the beginning was to completely replace ELF V1 using 
standard Digital products, not simply patch the current product and keep 
it alive.  The Digital product which provides the database is the 
Distributed Directory Service (DDS) component of MAILbus, a NaC/MIG 
product.  It is designed to allow queries in various forms, and manages 
the process of keeping current the multiple copies of the database around 
the network.

DDS does not currently provide a remote access user agent.  However, it 
was found VTX does provide the necessary network protocols, user 
interface, and architecture management features (failover control) which 
a large distributed application like ELF requires.  The project, then, 
consisted of developing software which connected VTX to DDS; creating the 
network environment in which to operate ELF; securing the sites and 
resources to provide ELF service; providing the employee data feeds from 
Europe, GIA, and USA; and providing the necessary documentation, user 
communication, and support structures.

While ELF V2 offers many functional, accessibility, and network 
performance  enhancements, it also has some acknowledged drawbacks: the 
VTX user interface may be unfamiliar to some employees, and a slower name 
searching algorithm within DDS which does not support name misspellings 
in the same way as did ELF V1.  Both of these can be addressed through 
user training and awareness.

A password access control system was developed for ELF V2.  Use of 
password access allows an employee to access ELF from any node and 
account to modify their ELF data.  With ELF V1, the employee had first to 
be registered in ELF by their system manager, and then had to access ELF 
from that specific account in order to modify their entry.  Whenever 
there was a problem which couldn't be handled by their system manager 
(e.g. node no longer available, registered erroneously to a cluster 
alias), the only recourse was the ELF Central Maintainer in DT.  By 
delegating password issuance to the local IS hotline function within each 
Digital facility (one or more sitecodes), local ELF support is now 
available to the individual employee.

Issues related to the transition from Version 1 to Version 2 have been 
addressed.  Any person properly registered in Version 1 will be able to 
assign themselves a password when they initially enter ELF V2.

The approximately 20 ELF remote infobase/servers have been provided by 
individual Digital facilities throughout the world.  These sites are 
supported via the ELF_SUPPORT Notes conference, by the geography hotlines 
and/or DT Operations Network Applications hotline, and finally by DT 
Engineering.

Implementation of ELF V2 within the Digital environment now will provide 
Engineering with valuable experience in improving our directory products 
thus enhancing our market position.

    As the original designer of the VTX form interface (but having nothing
    whatsoever to do with ELF in any of its versions), I should say a few
    things:
    
    1)  The interface was designed to exactly model the functions in the
    	public videotex systems.  In those it is important that the user
    	know when he is about to "commit" a transaction, as he may be
    	making plane reservations, etc.  That is what PF1-RETURN means.
    	Plain old RETURN means "next field".
    
    2)	If there are optional inputs to the ELF search function, which I assume
    	is what all the "/mumble=foo" is about, then these inputs should
    	have been made into additional fields on the form, with appropriate
    	fixed text explaining them.  Making the user type goofy strings
    	into the fields is a violation of the spirit of videotex user
    	interfaces.
    
    An additional point of interest was that at one time DEC actually had a
    product that *was* the software used by telephone companies for
    directory assistance.  It was called ASSIST-11.  It was written in
    MUMPS and ran on 11/70's.  (Unfortunately what we sold was the
    prototype, that was rushed out the door by product management.  As a
    result performance was not what it should have been.)

> Implementation of ELF V2 within the Digital environment now will provide 
> Engineering with valuable experience in improving our directory products 
> thus enhancing our market position.

From the comments here and in the ELF conference, I think we have provided
Engineering with valuable experience in improving our directory products,
WITHOUT the need to implement this disaster.

Bob

The user interface is only part of the problem.

I could live with the bad user interface if the thing WORKED.

It doesn't.

I can't find Jack Smith using his real name.  I have to know he's in the
database with his nickname.  If I search for John Smith I get a bunch of
them back, but not our V.P.  If I search for John F Smith (he is in there
that way, since that's his real name on his personnel form) the thing
times out.

ELF V2 doesn't work.

/john

    I prefer the new system, although I don't care for the fact that it's a
    Videotex application.
    
    Right now, in the current ELF, no one can find me.  You can't find me
    under Davis (maiden name) and you can't find me under Davis Hallyburton
    (married name).  You can only find me under Davis-Hallyburton or
    DavisHallyburton.  
    
    Under the new system, I can specify two given names and two surnames
    and the system will match on any combination: 
    
    Marge Davis
    Marge Davis Hallyburton
    Margaret Davis or
    Margaret Davis Hallyburton
    
    If you cannot find Jack Smith under the new ELF, it's probably because
    he's registered on PAWS as JACK, not JOHN.
    
    my .02,
    Marge

	Two part names has long been a known problem with ELF V1.

	BTW, Marge, FIND Marge Davis did not work for me with V2.
	FIND M DAVIS timed out.

	Interesting experiment. I looked up BOB REED and ROBERT REED.

	Phone book: 0 Bob Reed
		    3 Robert Reed

	ELF V2	    1 Bob Reed
		    3 Robert Reed

	ELF V1	    4 Bob Reed
		    5 Robert Reed

	Nice to see everything is consistant. :-) Anyone know which is
	correct?

			Alfred

I would suspect that the user interface *is* part of the problem.  VTX
was designed for folk that enter it once, then mess around for a while.
A "phone book" should be optimized for an immediate response.  If your're
on the same Ethernet as a server and your system isn't loaded, a normal
exact-match search should complete in less than two seconds.

We talk about application design and human factors usability to our
customers.  We have a large human factors group in Dec.  Were they
consulted when VTX/ELF was designed?

I'm not at all impressed by the decision to "use our own tools" -- that
smacks of the philosophy that, "if your only tool is a hammer, all
your problems look like nails."

Martin.

>    Under the new system, I can specify two given names and two surnames
>    and the system will match on any combination: 
>    
>    Marge Davis
>    Marge Davis Hallyburton
>    Margaret Davis or
>    Margaret Davis Hallyburton

In theory.  But it doesn't work.

Find marge davis

%ELF-E-NOTFOUND,  ELF entry not found

Find marge davis hallyburton

Common Name:   MARGE DAVIS HALLYBURTON 
DTN:  264-2847  Intrnl Mail Addr:  TTB1-6/D5  Org Unit:  FIELD SERVICE

Find margaret davis hallyburton

Common Name:   MARGE DAVIS HALLYBURTON 
Search Given Name:  MARGARET  DTN:  264-2847  Intrnl Mail Addr:  TTB1-6/D5
Org Unit:  FIELD SERVICE

Find margaret davis

%ELF-E-NOTFOUND,  ELF entry not found

Find marge hallyburton

%ELF-E-NOTFOUND,  ELF entry not found

Find margaret hallyburton

%ELF-E-NOTFOUND,  ELF entry not found

    Well, I just made the modify a couple days ago... it does say that it
    takes a few days.  I'm patient.
    
    grins,
    Marge

    I likewise have tried the new 'system' and find it tremendously
    lacking in it's UI and it's ability to retrieve useful information.
    
    I find it interesting that we are implementing the new ELF V2 system
    halfway between Halloween and Thanksgiving. Somewhat prosaic isn't it ? 
    
    /jim

    
    
    Prediction:
    
    o V2.0 will be implemented
    
    o some VP, Exec VP, or KO's secretary (or a combination of the above)
    will complain to same that V2.0 is the pits, that they can't find
    people for meetings, etc.
    
    o VP, exec VP, or KO will ask why 2.0 was implemented, etc.
    
    o discussion of why the elf "customers" were not listened to, etc.
    
    Summary:
    
    	- poor UI forced upon "customers" by time to market or other
        "goals"
    
        - mega complaints to head honchos
    
    	- lots of discussion, debate, etc. about why we should have
          done things differently.
    
        - lots of $ and productivity throughout the corporation wasted.
    
    	- lots of unhappy "customers"
    
    	- new UI on drawing boards soon after or return to V1.0 toute
    	  suite.
    
    too bad they can't sample the marketplace to find out what the
    real needs are... 
    
    -John

   I'm sure that ELF V2 will be able to meet the load requirements, nobody
will be using it...

   (I'd rather get ELF-F-NSA a few times than use the new one... ugh!)


					-mjg

(Now, if only the functionality were packaged in an RPC-able interface, and
we had an RPC, everyone could write their own interface... sigh.  I guess
that still doesn't fix the soundex problems tho'...)

    ELF V1 will be back on January 18th ...
    
    -- B

.50>I find it interesting that we are implementing the new ELF V2 system
.50>halfway between Halloween and Thanksgiving. Somewhat prosaic isn't it ? 
    
    Interesting.  The only more appropriate date would be April 1st.
    
.53>ELF V1 will be back on January 18th ...
    
    Did I neglect to note something in my black book?  Why Jan 18th?
    
    Jon

    We're not actually going to ask for MONEY for ELF V2. are we?

    Ford asked for money for the Edsel!
    
    Didn't mean they GOT any...

    I understand that ULTRIX users will not be able to access the new,
    unproved ELF.  If this is true, it gives the lie to our corporate
    message of supporting multi-platform users.
    

The responses to statements such as

	My name is R Jones.
	People can find me with ELF V1 by typing FIND R JONES.
	People cannot find me with ELF V2.
	Please respond to this BUG.

are

	This is not a bug.  We do not recommend using just a
	single initial.

This needs to be taken over the heads of the group which has approved the
implementation of ELF V2.

We should not be forced to use V2 until it has a searching capability which
is comparable to V1.

/john

    If I understand ELF V2 correctly, I could not find my supervisor, who
    does not use his given name.  Nor could I find my manager, for the same
    reason!  Not could I find any of a number of other people, unless I
    knew they weren't using their given names...
    
    By the way, if the developers were calling ELF V2 an improved user
    interface, I wonder if they know what they're talking about.  Clearly,
    the knowledge load on the user is increased.  Before, you typed what
    you thought the person's name was, and ELF V1 gave it a shot.  Now, you
    have to know the name exactly, right?  (If I knew the person's NAME, I
    wouldn't be looking it up in ELF, RIGHT? 8^)  And now do users also
    have to know about the incantation /G= /S= as well?
    
    Let's see... /G stands for FIRST, /S for LAST...?  No, GILBERT and
    SULLIVAN?  Maybe GOD, this STINKS?  It'll come to me... 8^)

As an ULTRIX user, I can do ELF V1 FINDs from the command line (someone 
reversed the ELF V1 protocol).  When V1 goes away, no more ELF for us
ULTRIX users.

                     
    	My biggest (of many) gripes is that when you type:
    
    find first last<return>
    
    not only does it ignore the command since you didn't enter PF1 Period or
    PF1 Enter or PF1 PF2 PF3 XYZZY or whatever the "return" key is, it
    doesn't erase the "faulty" command from the prompt!  It is not doing
    anything while it still looks like a perfectly valid command is after
    the last prompt on the screen.  No errors, warnings, no new prompt, no
    nothing.  Oh, you say, it moved the cursor location!  Why of course,
    that tells me it is ignoring my last line and is waiting for more
    input... ;-)
    
    	Enjoy it while it lasts -- ELF V3 will probably only accept punched
    cards submitted as batch jobs onto a standalone PDP-11.
    
    								-craig

>    	Enjoy it while it lasts -- ELF V3 will probably only accept punched
>    cards submitted as batch jobs onto a standalone PDP-11.
    
    Don't be silly.  ELF V3 will have the added enhancement that the
    miraculously quick lookups (wherein you'll also have to enter the
    middle initial of the person you're seeking; mandatory input) will be
    done from a datafile kept on a TK50.
    
    After realizing that ELF V2 is sadly NOT someone's idea of a joke, this
    scenario wouldn't surprise me...
    
    Jon
    

       
       
       	TK50? They wouldn't be that advanced! Each initial would be on a 
       TU58 on a 11/725, networked to a 11/730 with a DECtape. Of course,
       the "network" would be a null modem cable run between 2 DF01's. Did
       I forget to mention that someone has to place the phone on the
       coupler for each transaction?
       
       	Let's all send mail to SOCIAL::INVOLVEMENT. That's meant for 
       suggestions on how to save money.. Either that or we really
       have this code party.  Will the future developers of ELF V1+
       please stand up?
       
       							mike

    I hate to rain on this fun bashing parade, but I DO like V2's
    very nice habit of always being available. A cruddy interface is
    the price...
    
    Developers beware. NEVER take back capability. ELFV2's availability
    is sadly overshadowed by its reduced UI.

>While ELF V2 offers many functional, accessibility, and network 
>performance  enhancements, it also has some acknowledged drawbacks: the 
>VTX user interface may be unfamiliar to some employees, and a slower name 
>searching algorithm within DDS which does not support name misspellings 
>in the same way as did ELF V1.  Both of these can be addressed through 
>user training and awareness.
      ----------------------
    
    Please sign me up for the training that will help me spell things
    correctly.  I think it will take a few months so get me in one of
    the first courses started.  My family will be so glad to learn that
    I have finally learned to spell PERFECTLY 100% of the time. I had
    gone to school for 20 years from age five with-out getting to 95%
    correct.

    Can someone explain the details of the logic behind the above
    conclusion?  Sounds like F.M. to me...
       
    I think this interface should be added to the DECspell program.
    Instead of listing all the close matches for a wrongly spelled
    word just say "That word is incorrecly spelled, take another guess"
    and tease the users.
    
    

    Well they tell me it is easier to reach the server now... Well here
    in the Mill things are kinka sick on that count.  Well the good
    news is that it isn't working for the VP's Sec's too.  We may hear
    some complaints from the top soon.
    -----------------------------------------------------------------
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell
Unable to connect to site server of ASABET::"43="
Check vtx$server logical
$ el2 jon crowell

Common Name: JONATHAN CROWELL 
Search Surname: CROWELL  Search Given Name: JONATHAN, JONATHAN CLARK
Jon, PC Jon  DTN: 223-9757   Telephone:(508) 493-9757
Intrnl Mail Addr: MLO5-5/E71   Location: MLO5 pole 31D   Node: DIODE    
Username:  CROWELL   Org Unit: LOW END SYS/TECH, MicroVax EngineeringE

To:		Europe: Gabriel Barta @GEO GVA01::BARTA
		GIA: George Faucher @AKO   AKOV11::FAUCHER
		MEM: Lou Jacques @MRO      VLNVAX::LJACQUES
		NNE: Robert Yost @VRO      DISSRV::YOST
		STG: John Beirne @VRO      IAMOK::BEIRNE
		US: Bob Bradbury @PKO      FDCV06::BRADBURY


From:		Bill Kilgore               ESCROW::KILGORE
		Software Senior Engineer
		DECintact Development project leader
		DECtp Engineering	  


Re:		ELF V2


Gentle people:

I am writing this memo to voice my opinion of ELF V2. Your names were taken
from note 62.1 of the NOTES conference IAMOK::ELF, where you are mentioned
as representatives of the sponsoring groups for ELF V2. (The note is reproduced
below.)

I have seen many good software products and many bad software products in my
16+ years with Digital. With all regrets and due respect, I must place ELF V2
in the latter category. I will not enumerate my objections to the product
here; they have been voiced far better than I could express, in note 957.*
in the conference HUMAN::DIGITAL. I strongly encourage you to read note 957
and its replies in order to gauge the impact of ELF V2 on the Digital user
community. (The current tally is approximately 60 to 4 against the new ELF.)

I doubt very much that the problems with ELF V2 can be surmounted with "user
training and awareness", as implied in the management overview note below.
User training and awareness will not replace the ELF features that are lost
with V2. Nor will they make the awkward, slow and painfully verbose user
interface a less bitter pill to swallow.

Based on the reception of ELF V2 to date, I strongly urge you to delay the
replacement of ELF V1, continue the field test of ELF V2, perhaps in parallel
with V1 for better comparison, and take the reaction of your user community
to heart. Over the years, I have seen Digital turn slowly away from a "We know
what's best for you" attitude with its customers, to one of "Tell us what you
need, and we'll make it work, and work *well*." I sincerely hope that ELF V2
will not be foisted on a captive internal audience against overwhelming
objections, signaling a regression of attitude that cannot help but trickle
down to our customer base.


Bill Kilgore


--------------------------------------------------------------------------------

                <<< IAMOK::$1$DUA72:[NOTES$LIBRARY]ELF.NOTE;1 >>>
                         -< Employee Locator Facility >-
================================================================================
Note 62.1                >Management overview: the whys                   1 of 1
IAMOK::DEVIVO "Paul DeVivo @VRO, DIS/DT, 273-5166"   71 lines  12-OCT-1989 10:16
--------------------------------------------------------------------------------
                  ELF VERSION 2 - A MANAGEMENT OVERVIEW


This paper is designed to answer anticipated questions about ELF Version 
2 from a management point of view: why it is being introduced, and why it 
differs from ELF V1.

As most ELF users are aware, ELF Version 1, on the network since 1983, is 
not providing an adequate level of service to employees.  STG/Digital 
Telecommunications has developed ELF V2 to address these deficiencies.  
The project is sponsored by the Enterprise Network Applications Services 
Management Group (ENASMG) to which STG, MEM, and each geography has 
representation:

		Europe: Gabriel Barta @GEO
		GIA: George Faucher @AKO
		MEM: Lou Jacques @MRO
		NNE: Robert Yost @VRO
		STG: John Beirne @VRO
		US: Bob Bradbury @PKO

The objective since the beginning was to completely replace ELF V1 using 
standard Digital products, not simply patch the current product and keep 
it alive.  The Digital product which provides the database is the 
Distributed Directory Service (DDS) component of MAILbus, a NaC/MIG 
product.  It is designed to allow queries in various forms, and manages 
the process of keeping current the multiple copies of the database around 
the network.

DDS does not currently provide a remote access user agent.  However, it 
was found VTX does provide the necessary network protocols, user 
interface, and architecture management features (failover control) which 
a large distributed application like ELF requires.  The project, then, 
consisted of developing software which connected VTX to DDS; creating the 
network environment in which to operate ELF; securing the sites and 
resources to provide ELF service; providing the employee data feeds from 
Europe, GIA, and USA; and providing the necessary documentation, user 
communication, and support structures.

While ELF V2 offers many functional, accessibility, and network 
performance  enhancements, it also has some acknowledged drawbacks: the 
VTX user interface may be unfamiliar to some employees, and a slower name 
searching algorithm within DDS which does not support name misspellings 
in the same way as did ELF V1.  Both of these can be addressed through 
user training and awareness.

A password access control system was developed for ELF V2.  Use of 
password access allows an employee to access ELF from any node and 
account to modify their ELF data.  With ELF V1, the employee had first to 
be registered in ELF by their system manager, and then had to access ELF 
from that specific account in order to modify their entry.  Whenever 
there was a problem which couldn't be handled by their system manager 
(e.g. node no longer available, registered erroneously to a cluster 
alias), the only recourse was the ELF Central Maintainer in DT.  By 
delegating password issuance to the local IS hotline function within each 
Digital facility (one or more sitecodes), local ELF support is now 
available to the individual employee.

Issues related to the transition from Version 1 to Version 2 have been 
addressed.  Any person properly registered in Version 1 will be able to 
assign themselves a password when they initially enter ELF V2.

The approximately 20 ELF remote infobase/servers have been provided by 
individual Digital facilities throughout the world.  These sites are 
supported via the ELF_SUPPORT Notes conference, by the geography hotlines 
and/or DT Operations Network Applications hotline, and finally by DT 
Engineering.

Implementation of ELF V2 within the Digital environment now will provide 
Engineering with valuable experience in improving our directory products 
thus enhancing our market position.
    

    I, too, am writing to these people. I sugeest that you join in.
    
    All of you.

RE: .60
The replay focused on the question :
 	From now on ULTRIX user wil be without access to ELF v2.0 ?

My answer is (as VTX specialist):
Not for sure.  You have to install a DECwindow server (for ULTRIX !) 
on your VAXstation.  This way you'll have access to any 
DW client application (VTX TCP for DW) from your VAXStation.  

	Power of DW + VTX !

	_Giuseppe

Suggesting that access to ELF from Ultrix merely requires running a Decwindows
server on the Ultrix system begs the question.  Not all Ultrix systems run
DecWindows, and some of us still dial in at night from ancient steam-driven
VT52's (no, I haven't tried to run VTX from a VT52 lately).

Also, Ultrix has a very flexible subtask and message transport mechanism,
and many Ultrix folk are accustomed to building small "filter programs"
to convert information from one form to another.  This idea of using
small single-function programs in a larger structure -- as opposed to
large monolithic "do everything" programs, such as VTX/ELF is one of
the important advances of Ultrix.  Reaching ELF by adding yet another
large layer (DecWindows) is a poor solution.

Martin.

You should include Bob Igou (ELF Program Manager) as a recipient on anything
you write.

He called me yesterday.  One of the people working on the project sent a note I
had written in the ELF conference, where I had said that the lack of responsive-
ness from the ELF Team whenever someone reported a bug (that's not a bug, we
didn't want you to be able to enter just a first initial and last name) was
about to result in a grass-roots movement to raise the problem to the highest
levels in the company.

Naturally, he was concerned.  I hope he now takes a more involved role (starting
by reading the ELF conference, which he hadn't been doing).  Not surprisingly,
he was not very supportive of any plans to delay implementation.  It appears
that there are not any current plans to make the necessary major changes to DDS
which would restore the searching capability that ELF V1 had.

Their position is that ELF V1 had two shortcomings:  Poor server availability
and poor update propagation.  They believe to have fixed both of those, at the
cost of a degradation in the search capability.

I told him that it did appear that the server availability has been improved,
but that the update propagation hasn't been fixed (I would ask Marge Davis
Hallyburton to _immediately_ open a topic about her update problem in the ELF
conference).

But most importantly, ELF's ability to perform its primary function

	_FINDING_PEOPLE_

has been severely compromised by the poor searching capability.

In all our complaining, we need to focus on the bugs in the search algorithm,
since the user interface problems are truly secondary to the problems ELF V2
has performing its primary function.

/john

ELF V2 doesn't seem to like middle names much either

>             <<< Note 957.71 by COVERT::COVERT "John R. Covert" >>>

take John's name here as an example.

try 
$ find john * covert    (this is using ELF V2, BTW)

and you get 
%ELF-E-INVLDWILD,  Invalid wildcard

This worked in ELF V1.
You have to remember that a middle name is a given name, and as such
you should probably do :  $ find john* covert

But of course, this takes ages, and you only get peoples middle names/initials
if they've entered them into the search fields AND you do a /FULL search.

I've noticed a lot of entries in ELF V1 that had middle initials. A lot of
people (not myself, but never mind) like to put their middle initial in.
However, ELF V2 seems to make a mess of it! Are we all having the way were 
are called changed because of ELF V2? :-)

Can anyone answer this one too ...
If (say) Richard Bloggs enters his nickname in ELF V2 as 'Dick' and this
then propogates to the phone book, what do I use to send mail to him?

If the phone book lists him as 'Dick Bloggs @ XYZ' then do I send mail
to 1) 'Richard Bloggs@xyz' or 2) 'Dick Bloggs@xyz'

	Craig.

    Maybe we ought to just publish a phone book?  You know you can't
    trust a computer.
    ed
    

    Boy, are the developers of ELF V2 in for a big surprise when EVERYONE
         starts using VTX ELF instead of ELF F "X".
    
    Until I read this set of notes, I thought the reason that I wasn't
    finding anyone using the VTX version was because they hadn't been
    loaded into the database.  Now I understand it's because I didn't have
    the EXACT name.  You can imagine how ticked I was to find this out.
    
    I'm especially mad now, because no one will find me in the VTX version
    unless they know my FULL first name.  And there are two of us here in
    the company with a different first name.
    
    I agree with several of the previous replies.  No new version of
    anything should EVER PROVIDE LESS FUNCTIONALITY than previous versions.
    The User Interface arguement is a smokescreen that obfuscates this
    far more important issue.

>           <<< Note 957.73 by WORDS::BADGER "One Happy camper ;-)" >>>
>                       -< back in the good old days.... >-
>
>    Maybe we ought to just publish a phone book?  You know you can't
>    trust a computer.
>    ed
    
    You know, if net mail addresses were printed in the paper phone book, a
    lot of need for ELF V* would be obviated, much of the time.  How many
    times do you use ELF to look up a phone number?  How many time to find
    a net mail address?  I know for me, the latter exceeds the former...

    Re: "The User Interface arguement is a smokescreen that obfuscates . . ."

    The User Interface is a smokescreen that obfuscates . . .

    So, there are at least two fatal problems with this piece of work:

	I can't use it

	I don't want to use it

I think the user interface issue is a major part of the problem -- it it's
hard to use, people won't use it.  (I include search algorithms, speed,
 and command line interfaces, so we're sort of talking about the same thing.)

A few notes:

-- VTX is now becoming unavailable: is it dimensioned to handle the additional
   ELF traffic?

-- The database doesn't handle "multinational" chacters.  It won't let
   me search for "G�ran Carlsson", and changes the � to a random 7-bit
   character when it updates the display.

-- I can reliably hang my process by typing the wrong command.

-- Soundex (well, we could do better) is essential for people looking
   for names in a foreign language.

I'll post a longer message later.

Martin.

    Sounds like a solid 'design by committee' has replaced a 'back room
    hack'.   Bring back the hack!
    

I've mailed off my views on the new ELF, I could not however copy Bob Igou 
because everytime I tried to use ELF V2 to find his location I got;

	That information is not available.

Meaning ELF V2 was unavailable at that time, so much for the so called 
improvements in availability (this happened yesterday as well).

Dave.

    Send mail to Bob at IAMOK::IGOU.

    I think a VTX client for Ultrix is on the list of stuff being worked on
    by the VTX development group for the next version.  PC's too.
    
    The VTX code itself is easily capable of handling the load of being
    used for ELF throughout the corporation.  That is exactly what it was
    designed for.  But the actual database searching application, which
    hangs off the back of VTX, is more likely to be the bottleneck.
    
    Of course, the machines running the public VTX servers have to be up,
    and the network has to stay connected.

    I must first admit that I didn't check thru the 81 responses here
    before mine, but I have a question for other Workstation users:
    Does VTX ELF do "quirky" things to your VAXstation?  I'm a
    non-technoid, and do not understand why my windows all change color,
    my screens flash, and finally VTX ELF does come up.  The only way
    I can reset everything is to reboot my VAXstation.
    
    Suggestions?  Reasons?  Comments?  
    
    Thank you!

    Based upon my experience building VTX databases that used color, you
    can expect the color attributes on your monitor to get messed up when
    you exit from vtx.
    
    Ken

    Do a SET TERM/NOREGIS before invoking VTX.

    Try upgrading to VTX Version 4.0.  It should leave colors alone.
    _bill

    Or the current EFT version of DECWINDOWS,

I just looked and I cannot find my Software Tools Review Board notes from
the early eighties when we (STRB) funded the original development work on
ELF. Unfortunately I've either misplaced or discarded them. Too bad. If I
could find them now I'd be able to quote some of the design goals of the
original prototype which were upheld in the V1 product (and which aren't
even addressed in the V2 product!) I know some of these included the
ability to locate approximate matches. Hell! That's what the thing was
really for - finding someone when you DIDN'T know how to spell their
name! I believe John Hrones was the original developer. I'll contact him
and see if he can dig something up.

BTW, I agree with Walt Lamia a few back - I use the phone book for DTN's
but ELF for Nodename::Username.

-Jack

PS. BTW, take a look at 22.6 in IAMOK::ELF for a good laugh.

    re .63

>   TK50? They wouldn't be that advanced! Each initial would be on a 
>   TU58 on a 11/725, networked to a 11/730 with a DECtape. Of course, the
>   "network" would be a null modem cable run between 2 DF01's. Did I forget
>   to mention that someone has to place the phone on the coupler for each
>   transaction?
       
    I too will admit that I was first beguiled by the "availability" factor.
    That has, however, worn away and I now see subtle, yet distinctive and
    inmistakable, signs of a hammer looking for nails.

    We could, of course go back to use that famous system, now of folklore,
    that heralded but now forgotten missing link to the VAX: the LAX-11,
    the Longword Address eXtended [PDP]-11, the extended memory version of
    which was better known, of course, as the EX/LAX-11. And let us not forget
    PTOS-11, the Paper Tape Operating System. All you needed to run it was
    a wall outlet and a warehouse with the appropriate cubic foot capacity,
    Empire State Building size being the supported functional minimal
    requirement.  :-)


    /pete[r][s] ... oops, only "PETERS" now works, not PETE, PETER, or PETES

          |
          V
    /Peters <----- Yo, THAT'S A "S"
          ^
          !

    P.S.  The implementation team has my full support. I don't think it was 
    they who decided that users just need training to use reduced 
    functionality effectively (i.e., not try things that used to work).

    In this era of cost-savings, perhaps to save on Digital's internal use 
    of disk space we should all be retrained to use shorter one and
    two-syllable words only. I'm sure that with the proper education we
    would quickly forget the need for longer words. For example, using DEC
    everywhere in place of Di-gi-tal and dropping the DEC prefix from
    product names ("DEC-Win-dows" would not be allowed under the new
    guidelines) alone would save:

        3-4 bytes * 120,000 or 125,000 (or is it 140,000?) employees * 500
        occurrences of "Digital" in one's personal files (mail,
        documentation, etc.) -- I counted -- then probably add quintuple
        that at least for all official occurences of "Digital" and
        "DECproduct", we get 1,680,000,000 bytes of storage right there.

    From my experience, and I wasn't much of an ELF V1 user, what
    it returned on a did not find, but did find
    didn't help that often.  Examples - seaching for
    DAVE PALMER spelling his name PALMBER or PALLMER won't find
    him.  If I truely didn't know how to spell that name, I wouldn't
    know how to find him in ELF V1.
    
    However, FIND PAL* in ELF V2 finds him.
    
    Another example, what may be the ELF V1 equivelent of a 
    FIND PA* LE* @PKO returns some, none starting with PA, none in PKO.  
    ELF V2, if it doesn't time out, returns the correct information.
    
    So I toss out if you really don't know the last name, and what you
    type in is "correct" and shorter than the target name, ELF V2 will
    find it when it doesn't time out (and none of my searches of this
    type have yet).  On the otherhand, if there is a  name
    "in between" what you type and the target name, ELF V1 will never find
    the target name, as far as I know.
    
    As for what happens when ELF V2 doesn't find the one you want, the
    way to continue on is with a /g swith or perhaps a /s swith or both
    or some other additional information.  This also means the person
    has to add how they want to be called, or change it in the proper ways.
    Even here, I suggest that ELF V2 can search MORE than was ever possible
    with ELF V1.  Marge may be an example here when she has her entry set
    up.
    
    And a PF1 enter is not required.  On systems that use the VTX command
    as supplied by the kit, VTX/DEV=A1 lets you use ELF without ever
    hitting the PF1 key.
    
    And re: the note having trouble connecting with el2  (name).
    I suspect you'll find EL2 to be defined to execute ELF_FIND.
    THIS IS NOT PART OF ELF V2.  Using the ELF keyname in VTX is more ELF
    V2.  If THAT has trouble connecting, when implemented, you should
    contact your VTX and/or ELF support.
    
    RE: is this what we sell.  VTX is 29th on the list of software
    revenue (out of 250 or something like that).  That doesn't include
    the hardware or service revenue it generates.  And yes, ELF V2 is
    being offered to ASSETS.
    	_bill

re .87:

>                                             That's what the thing was
>really for - finding someone when you DIDN'T know how to spell their
>name! I believe John Hrones was the original developer.

How appropriate!

    The only thing I ever use ELF for is to find out where I can send
    electronic mail to someone (assuming I can find an ELF server...);
    otherwise I'd be willing to look in the phone book for all the
    reasonable permutations of the person's name, since it nearly always
    takes a lot of tries to get a connection to a server around here.

    
    I wonder if a "hypercard" type hyperinformation application wouldn't be
    more appropriate as an alternative to the current ELF application. 
    125,000+ file folders, one for each employee, containing LOTS of
    information (DTNs, outside telephones, office addresses, home
    telephones [if the employee wanted it listed], job function, skills
    [enabling people to find expertise when needed], group, mailstop,
    poles, and on and on).
    
    

	Of course ELF V2 is not all bad. There are, if you get used
	to the user interface, some nice things that you can do.

	For example, you can add names to your record for people to
	search for. Try FIND EASYNOTES and you'll get me. I've also
	modified the organization section (you can add three items
	there) to show that EASYNOTES.LIS is a spare time activity.

	This could be even more useful if people want to play around
	with it. For example if all the moderators of this conference
	added MODERATOR as a surname and DIGITAL as a given name than
	FIND DIGITAL MODERATOR would, I assume, give you a list of
	the moderators. Now if all moderators did this and had the
	additional search surname of MODERATOR we'd give the search
	algorithm some serious testing in duplicate keys. :-)

	I'm not sure this is/was an intended use of this functionality
	but it might solve a lot of problems. Or cause some. :-)

			Alfred

>>           <<< Note 957.63 by STAR::MFOLEY "Rebel without a Clue" >>>
>>...       
>>       Either that or we really
>>       have this code party.  Will the future developers of ELF V1+
>>       please stand up?

I'll be more than happy to contribute to the DECwindows UI.  I'm sure it
wouldn't take me more than a day for that part (two at the most).  Of course,
I'd still usually use a command line interface, but DECwindows would be
a nice option.

-steve-

    One of the problems that ELF V2 is supposed to have solved is
    propagation of information in a timely manner. Well it was days
    ago that Marge Davis Hallyburton added DAVIS as an alternate
    surname yet I still can't get FIND MARGE DAVIS to find her. I've
    been told that FIND EASYNOTES doesn't work (to find me) for everyone
    either. Looks like that problem may still exist.

    		Alfred

    Just checking, but is she sure she gave the modified request
    correctly?  I can see my changes on the local server in two
    or three minutes.  Haven't checked to see how fast the other
    databases get updated though.

    Well, I have just seen this turkey for the first time, and wearing my
    hat as original designer of VTX, including its much-maligned forms
    package, all I can say is that ELF V2 is the *worst* use of videotex
    I have ever seen.  The implementors have completely missed the point
    of how videotex applications should look, and tried to make it look
    like any old VMS program.  "ELF>" prompts?  Commands entered into
    fields?  Come on!
    
    I was able to use the French PTT directory system several years ago,
    and it is *much* more sophisticated than this.  And easier to use.
    Even if you don't speak French.  Better search algorithms too.

    Hehe, you should see the bastardisation called VTX/DECWINDOWS.

    re .97:
    
    Yes, I gave the Modify correctly.  I've called my ELF Administrator
    in MKO to check up on the change.  So far, to no avail.
    
    Marge

RE: .99

	What a joke!! A truly poor implementation of a DECwindows interface.

								mike

    just for information sake, there is a petition going around here
    requesting a delay in the official turning on of the new elf and the
    shutting down of the old version.

    Lou

Just for giggles, I entered ELF V2 for the first time.

I then told it to change my password.

Good grief, I am not on the right node, I should use
ELF 1.4 or contact my ELF administrator before It will do it's thing for me.

Now if I can remember WHICH node that used to be in this cluster, or is it still
here, or, let me see, hummm ...



		Oh, I digress }8-)}

					Bill

Where is this discussion supposed to be leading?

Does someone believe we can stop the implementation of ELF V2?  Or is this
intended to be a general bitch session?

When an application is constrained to use the truly junky user interface
provided by VTX and the doggie database search and update capabilities of
DDS, bitching at the application is probably a case of misdirected anger
(unless you can home in on the people that set up those original constraints).

Obviously, if there is as high a degree of dissatisfaction with the product
as seems to be represented here, people should be talking to their functional
(GIA, EUR, MEM, SSMI, CIS, etc.) representatives to try to delay the final
implementation.  But where were all the folks who are now dissatisfied when
the product was being field tested?  Was there a significant problem with
the way that the field test was administered that needs to be elevated to
DIS management?  Or was everyone who is now unhappy just being lazy about
trying the product out then?

The problems need to be prioritized.  A lot of the user interface stuff can
probably be gotten around with shell procedures of some sort.  The fact
that reasonable name searches time out sounds much more ominous, however.

So what's the next step (or is there one)?
								paul

    Can I ask a dumb question?  What good is ELF, old or new?  I mean,
    I can look up a phone number and location code in the telephone
    book far faster than I can do it using any of the existing permutations
    of ELF.  If you use MTS (or MR or whatever it's called these days)
    to send mail, why do you need to know the destination node address?
    Or why would you want to? I thought that was the whole point.  Or am 
    I missing something?
    
    Al-An_ex-tecchie_who_relies_mostly_on_All-in-1
    

I'm an Information Services Manager responsible for delivering ELF V2 service
and support to the end-users within the CMAS (Marlborough) geography. I agree 
that it will be a big mistake to implement ELF V2 in its current state.

My understanding is that a vote will be taken by the ENASMG Committee (see 
below) tomorrow to decide about whether to implement ELF V2 at this time. One
of the alternatives to be considered is to implement ELF V2, but continue to
make ELF V1 available. 

I recommend that those of you who feel strongly about this, please send mail 
to your committee representative as soon as possible.

Jim

-------------------------------------------------------------------------------
The ELF V2 project is sponsored by the Enterprise Network Applications Services 
Management Group (ENASMG) to which STG, MEM, and each geography has 
representation:

		Europe: Gabriel Barta @GEO GVA01::BARTA
		GIA: George Faucher @AKO   AKOV11::FAUCHER
		MEM: Lou Jacques @MRO      VLNVAX::LJACQUES
		NNE: Robert Yost @VRO      IAMOK::YOST
		STG: John Beirne @VRO      IAMOK::BEIRNE
		US: Bob Bradbury @PKO      FDCV06::BRADBURY

       RE: .106
       
       	Us techno-weenies up here in ZK don't have a clue as to how
       ALL-IN-1 (correct trademark) and MTS and MR work and to tell
       you the truth, WE DON'T CARE.  (well, most of us anyway) We like 
       our VMSmail and NMail thank you very much. It works and we don't
       want the additional overhead of ALL-IN-1.
       
       							mike

    The advantage of ELF over the Phonebook is a) it's more current and b)
    we could correct mistakes (in V1 at least).
    
    - Andy ��� Leslie
      VMS CSSE Newbury

    re: .107 Mail sent.

�        <<< Note 957.100 by SCARY::M_DAVIS "Marge Davis Hallyburton" >>>
�    Yes, I gave the Modify correctly.  I've called my ELF Administrator
�    in MKO to check up on the change.  So far, to no avail.
    
    Have you tried to connect to other VTX servers and then give your
    request?
    
    Try these, do 
       $ DEFINE VTX$SERVER COBRA::"""43="""
    or try some of the other servers,
       "VTX$SERVER" = "RUTLND::"43=""
       "VTX$SERVER02" = "DECNA::"VTX=""

    re -.1
    
    The "Marge Davis Hallyburton" epic should really be taken offline here
    and brought to the IAMOK::ELF notesconference. The advice given in
    -.1 demonstrates the difference between the DIGITAL and the ELF
    notesconference - the technical advice provided in the ELF notesfile
    is accurate (or at least claims to be).
    
    Rumi Gonda's advice isn't correct - it doesn't matter which VTX server
    Marge used to get to ELF - she got to it. The issue is that after she
    got to the ELF server, either  ELF itself, or DDS (which handles the
    database and its updates) didn't work her transaction. Issues of this
    type should go into the technical support notesfile.
    
    I don't think that 957.* should be taken up with technical support
    issues, but rather as a forum discussing the implementation plans of
    ELF V2. I only recently learned of this discussion (saw enough flames
    in IAMOK::ELF so that this wasn't an eye-opener), and I've enjoyed not
    seeing each comment handled by a "You're not using the tool correctly -
    it'll work if you say F/S="BLOGGS"/G="ROBERT"" or some other claim that
    the user interface isn't being used correctly.
    
    Today's ENASMG vote may postpone the implementation of ELF V2,
    and continue the support (for the time being) of V1. If it doesn't,
    what will YOU do? Will you go back to using the phone book and your own
    rolodexes (electronic or otherwise)?  Or will you use the new ELF and
    hate it?  Is there another option I'm forgetting about?
    
    jack
    
    

    re .111 and .112
    112 is correct, please DON't follow the advice in .111.  Nothing
    will give worse results than clogging up a single server.  In this
    case, one where the problem(s) were, anyway.
    There was a technical issue or two, now resolved with the Marge Davis 
    update.
    
    	There is another option.  MOST of the comments in all notesfiles
    boil down to "easily" :-) settable defaults.  Most of the searches
    mentioned (except Jon Arnold's) return in just a bit more time under
    "the best of conditions".  One other option is letting the searches
    go a bit longer, and more experience with the products so the time
    needed can be cut down.
    _bill

    Thanks to all for their suggestions.  I checked today and I can find
    myself under any of the following combinations:  
    
    Marge Daivs
    Marge Hallyburton
    Marge Davis Hallyburton
    Margaret Davis
    Margaret Hallyburton or
    Margaret Davis Hallyburton
    
    It took longer than the advertised "2 days" to make the update, but
    once it was made, it worked.  Under V1, I am still only available
    under either DavisHallyburton or Davis-Hallyburton, neither of which is
    my name.
    
    signing off this discussion now,
    Marge

>    Today's ENASMG vote may postpone the implementation of ELF V2,
>    and continue the support (for the time being) of V1. If it doesn't,
>    what will YOU do? Will you go back to using the phone book and your own
>    rolodexes (electronic or otherwise)?  Or will you use the new ELF and
>    hate it?  Is there another option I'm forgetting about?

    There is nothing wrong with implementation of ELF V2 as long as that
    implementation does not included the stopping of feeds to ELF V1. In
    fact giving ELF V2 a serious field test with real loads is a good idea.
    Shutting down something that works is not a good idea. In fact as
    people use V2 more for the cases that it does work for, the load on
    V1 will make things easier for people who still need V1 features.

    Oh yes you have forgotten several options. One is to grow ones own ELF.
    I would say that that is a real possibility if the user interface and
    search algorithms are not fixed in V2 before V1 is shut down. An other
    option is to start up V1 servers on non-DIS nodes. If direct feeds are
    not provided, which will show that there is little real management
    support for efficient, user friendly tools, than a way will be found 
    to get a feed from V2. Never under estimate the obstinacy of the
    engineering organizations. :-)

    			Alfred

    re: .106  (What good is ELF...?)
    
    ELF is also used by SWS Residents.  Residents around here are not
    given Digital phone books.  Even if they were, they would be prohibited
    from taking the Digital Internal Use phone book on-site with them,
    unless they intended to guard it all the time (no thanks...).
    
    ELF is about the easiest way to get a node name or phone number,
    especially when its 5:35 PM, the switchboard is closed, your unit
    secretary has gone, but you know that the person you need is likely
    to be at his/her desk.
    
    FWIW
    
    -- Russ

re .106:

Here in the UK (yes, me again!) we have a phonebook for the UK, a phonebook
for the USA, and possibly a phonebook for the rest of Europe (though I can't
say, never having seen one). That just leaves GIA, and the people who came
(or left or moved or got married) since the last printing *that got through
to my office*.

ELF (if properly implemented - the "ideal" ELF) offers the following advantages:

	(1) Always up to date within 1 day (1 week at most).

	(2) Tracks moves or marriages (or other name changes).

	(3) Identifies people, through the free text field (RIP). I find it
	    really useful to find out that someone is a CIM marketing manager,
	    or is responsible for public relations in Italy.

	(4) Now there's an idea! How about searches on the free text field, so
	    you could ask ELF "who's the corporate DECfoobar product manager?"
	    and get name, DTN, location and VAXmail...

Suggesting recourse to the phonebook is a counsel of defeat. Sure, in the
present circumstances it may be the only thing to do, but it's not really
where we (the distributed computer system, enterprise integration people,
remember?) want to go.

/Tom

re .117:

Yes, there is a European phonebook called "European Communication 
Directory". They are distributed to most locations as a general delivery to 
the mailroom. If you don't find out there's a new issue and grab one fast 
you loose out. My last copy is Nov'88, I requested one from the 
distributors but nothing showed. Thus, I rely on ELF.

Don't forget a good ELF implementation would save trees because we would 
not need phonebooks.

Dave.

    re: .116
    
    I was a SWS resident for over 5 years, so I know about off-hours
    access.  I find that is the only time ELF is usable anyway, at least
    the V1 permutation.
    
    There is absolutely NO reason why you should be denied a phonebook.
    Digital phonebooks are marked Internal Use Only.  So is Sales Update.
    The guidelines for handling material marked this way clearly indicate
    that you are not handling State secrets.  You are required to use a little
    common sense in handling the material - that is basically all. 
    Check the PP&P for the specific guidelines.
    
    I have heard of some overzealous admin types making up local rules
    which forbid the removal of phonebooks from company property.  Those
    people are idiots and their rules are idiotic.  Don't let some cretin 
    stand in the way of you doing your job; generating revenue is
    important, silly rules are not.
    
    re: .117
    
    Not everyone is in Engineering and sitting in front of a terminal
    or workstation all day.  Managers, for example, spend a lot of time
    on the phone, in meetings and such.  Tell me which is better when
    I need a number: opening my top left desk drawer and reaching for
    the phonebook, or taking over a minute to log on (assuming my secretary
    isn't in my account, locking me out) and fumbling with ELF or whatever?
    The phonebook wins hands down, plus it's portable.
    
    There is only one benefit of ELF to me; namely, finding a phone number
    in those rare instances when I have access to my account on the
    Digital network but not to a phone book.
    
    I like your other ideas, but unless you can duplicate the convenience
    of the phonebook, you won't be able to replace it.
    
    re: .118
    
    A properly implemented ELF system won't save all that many trees.
    Don't labor under the illusion that it will.  The idea is to use
    technology to solve business problems, not create them.
    
    re: a few back
    
    It's interesting to note that "techno-weenies" don't use the products 
    they engineer for our customers.  I see that as a problem.  Don't
    get me wrong; I think of VMSmail has having a definite niche.  But
    considering the size of this corporation, the diversity of its
    employees and the investments we have made in a corporate E-mail
    infrastructure, I can't see why one needs two systems.  Either MTS
    doesn't work, or people are wasting time chasing information they
    don't need.  Neither explantion makes a whole lot of sense to me.
    
    Al
    
    

>    I have heard of some overzealous admin types making up local rules
>    which forbid the removal of phonebooks from company property.  Those
>    people are idiots and their rules are idiotic.

Though not on the May 1989 phonebook, the July 1988 phonebook sez:

	For Digital Internal Use Only
	Digital Equipment Corporation
	Confidential and Proprietary
	Do Not Remove From Digital Property

on the front cover.

       RE: .119 (technoweenies)
       
       	For many of us IN Engineering, ALL-IN-1 and assorted stuff doesn't
       fit with the way we, the development groups, do our business. I 
       believe that there are PLENTY of organizations in DEC that can
       really use ALL-IN-1 and use it better than us. We should be looking
       to those people for the feedback. And we should listen if we don't
       now. Meanwhile, we have a product to put out (and believe us, we
       are VERY busy) and ALL-IN-1/MTS/MR just doesn't fit our business
       plan. At least right now. In the meantime, we need an excellent ELF. 
       Right now, we don't have one and it doesn't look like we're gonna get 
       one unless we do something about it. 
       
       						mike
       						VMS Development

Re .120:

>	Do Not Remove From Digital Property

Aw, that just means you have to store the phone book at home on top of your
loaner terminal.
				/AHM

    RE: .120

    Yes, they used to say that.  How quickly I'd forgotten.  "Internal
    Use Only" AND "Confidential and Proprietary".  Which one was it?
    "Confidential and Proprietary" is not a real classification, and
    the words and actions used to expand upon "Internal Use Only"
    contradicted the official classification.  I used to ignore it just
    the same, with support from my management.  It's called doing what's 
    right.  I guess some else realized what phone books are used for
    and had the contradictory language removed from the cover.
    
    Speaking of stupid rules, I recently had an admin type inform my
    secretary that badge numbers were confidential, and thus my list
    of employees and their badge numbers that hangs on the wall next to 
    my desk was "illegal".  Imagine that. Everyone is supposed to be
    walking around the building with their name and badge number clearly
    displayed on their person, yet that information is at the same time
    "confidential".  I told her that she could do with her list as she
    wishes, mine is staying where it is.  There are rule and then there are
    "roolz" I guess....
    
    re: .121
    
    No one is suggesting that you use ALL-IN-1 as your user mail agent.
    My understanding of how MTS is supposed to work (and it appears
    that my understanding may be incorrect from a quick perusal of the
    MTS section of the phonebook) is that you address your mail to a
    person at a location code, and MTS figures out both which system
    that person is on and what user agent they use (e.g., ALL-IN-1, VMSmail,
    or printer).
    
    That means if I want to send mail to Mike Foley @ZKO, I don't need
    to know that his electronic address is STAR::MFOLEY.  If I were
    a VMSmail user I would simply address it to MTS$::"ZKO::MIKE FOLEY".
    I suspect, however, that this technique may not work in all cases 
    since: 1. engineers probably hop from machine to machine on a
    regular basis and 2. MTS mail is not currently handling VMSmail
    addresses (but it should).
    
    Al
    

    
Re .120:

>	Do Not Remove From Digital Property

    Or, if you are in the field, keep it in your company car }8-)} .
    
    						Bill
    

Re: < Note 957.123 by HOCUS::KOZAKIEWICZ "Shoes for industry" >

>    That means if I want to send mail to Mike Foley @ZKO, I don't need
>    to know that his electronic address is STAR::MFOLEY.  If I were
>    a VMSmail user I would simply address it to MTS$::"ZKO::MIKE FOLEY".
>    I suspect, however, that this technique may not work in all cases 
>    since: 1. engineers probably hop from machine to machine on a
>    regular basis and 2. MTS mail is not currently handling VMSmail
>    addresses (but it should).
    
Well, your last observation sure is right. As an experiment, I tried
to send myself a message using this address string in VMS mail:

    To: mts$::"ZKO::CHUCK MURRAY"

The message got sent to Wanfang Murray! (Wanfang is a software engineer
at ZKO and she's also a friend of mine, but she's definitely not "Chuck
Murray.")    

   Watch out with this MTS/ALL-IN-1 vs. VAXmail stuff, or someone in
corporate might try to phase out VAXmail next...


					-mjg

    Back to our regularly scheduled topic, I understand that at the ENASMG
    meeting earlier this week, it was voted 5-1 to go ahead with the ELF V2
    implementation on the scheduled date.  The one dissenting vote tried to
    exercise the "veto power" supposedly inherent in each of those members,
    but was told that for this topic, veto was not allowed.
    
    Doncha just love politics...
    
    Jon

re .123:

On your second point, Al, that may be what MTS _should_ do, but where I live
it certainly _doesn't_.

I've never inhabited ALL_IN_1, because it takes me so long to get anything
done (no blame, just my experience). So I always set it up to send mail to
me on VMSmail, which I then set up with a forward to my workstation. On the
workstation, I set up a logical, like 'MR$:==nm%HAMPS::MRGATE::' and then
I send mail to MR$"ZKO::MIKE FOLEY". So far so good, except that I have to
set up the logical.

But once it gets to the ALL_IN_1 system at ZKO (if there _is_ one {8-) ),
then unless Mike has an ALL_IN_1 account _and_ has done all the things that
I did at my end, he won't get the mail. At best he will get some paper mail
printed out on some printer somewhere, and subsequently adding to the fire risk
of that towering pile of unclaimed documents in the printer room.

But that's the best-case scenario! When I was at DECworld, the system manager
there set up ALL_IN_1 so that you would have had to send to "DWL::WELSH.TOM".
Yes, that's right, a period! Of course, in France they use periods instead
of commas and vice versa quite a lot (plays hell with your decimal arithmetic).
The US equivalent would be "DWL::WELSH,TOM".

My point is - how on earth does anyone in America know whether to mail me as

		BST::Tom Welsh

		    or

		BST::Welsh,Tom

		    or

		BST::Welsh.Tom

		    or

		BST::Welsh

		    or

		something else even weirder?

Now, when it takes as much as _days_ for the message to arrive, not be
recognised, and sent back again by that old POSTMASTER (how I hate him - he
never seems to do anything except reject my perfectly legitimate mail), there
is a limited number of these cycles you can go through trying to get mail
through to someone.

I'm sometimes tempted to send a boy with a forked stick...

/Tom

    Am I the only one who sees something amiss with a mail system that
    serves 120,000 employees (give or take) and requires substantial
    numbers of users to know network addresses to use it?  That no matter
    how fine an ELF system you devise to supply people with those network
    addresses, you are attending the symptom and not the problem?
    
    I recognize that MTS doesn't work.  Are we better off working around
    it or fixing it?  
    
    Al
    

    I agree with .-1.  In fact, the predecessor to MTS and mail or MTS and
    All-In-1, EMS, written in MUMPS-11 and running on PDP11s had more
    capability in terms of determining "which Joe Smith" than MTS provides
    today. (Granted every employee in the company didn't have access and
    there was a relatively small number of EMS nodes.)  But, when you
    specified "Joe Smith" as an addressee, and there was more than one
    registered in EMS, you got a list to choose from showing each Joe Smith
    and location.
    
    So, .-1 is correct in that a separate ELF access shouldn't be required
    to send mail.  ELF is still useful as a telephone directory because the
    phone book only includes US and Canada and it is not published
    frequently enough to keep up with people transferring or facilities
    moving.

    We're fixing it.

    The implementation of ELF V2 is an emotional event because it provides
    evidence that the forces of bureaucracy and authority have crippled
    our ability to deliver quality solutions.  It is a microcosm of
    Digital's crisis as a corporation.
    
    
    MATT

>< Note 957.129 by POCUS::KOZAKIEWICZ "Shoes for industry" >
>    Am I the only one who sees something amiss with a mail system that
>    serves 120,000 employees (give or take) and requires substantial
>    numbers of users to know network addresses to use it?  That no matter
>    how fine an ELF system you devise to supply people with those network
>    addresses, you are attending the symptom and not the problem?

Well, as long as people have the silly freedom to choose the same name 
as somebody else who already works here, (don't all those Joe Smiths ever get
mad at one another for stealing their names?) and various people have different
distinguishing characteristics (real first names, nicknames, site codes, 
etc.), it seems that some manual step for pre-qualifying your addressee is 
not out of the question.  You have to know SOMETHING that makes your
correspondent unique, and he or she may not always be available to tell
you what it is.  Right now our major conventions for making the distinction
are node name and site code, relative constants to indicate the business
need for which we are contacting a given person (if either changes, maybe
that means our reason for contacting that person is no longer appropriate).

Barring the "read sender's mind" prompt for a mail system, what improvement
would you suggest?

- tom powers]

A long series of comments on Elf Version 2 that was entered into
IAMOK::ELF over the weekend brought up a few that seem relevent
for "how we work at Digital."

-- A visually-impared employee pointed out that the new system is
   essentially unusable from her Braille-terminal, while the previous
   version "is easy to use (when there is a server available)."
   Note that this employee cannot use the paper phone book.  (Note 91.1)

-- A system manager pointed out that the employee-settable passwords
   are stored in the database in human-readable text.  While the
   file is not directly readable by a malicious individual, were
   one to gain access, it is likely that real system passwords
   could be compromised. (Note 91.101).

-- One of Dec's Human Factors experts, Betsy Comstock, posted a
   succinct course of application interface design (note 91.76) that
   would be useful reading for anybody designing a similar system.

Martin.

    < Note 957.134 by REGENT::POWERS >

>Right now our major conventions for making the distinction
>are node name and site code, relative constants to indicate the business
>need for which we are contacting a given person (if either changes, maybe
>that means our reason for contacting that person is no longer appropriate).
    
    Only in certain organizations is this true.  How many organizations,
    discounting Engineering, rely primarily on MTS to deliver E-mail?  I
    honestly don't know the answer, but I'd be mighty suprised if it
    wasn't most of them.  Most people would then need to rely on
    name/site-code, not node. 

>Barring the "read sender's mind" prompt for a mail system, what improvement
>would you suggest?
    
    1.  Make MTS work the way it should.
    
    2.  You will always have name conflicts, but let's be real.  How
    many name/site code conflicts are there in the corporation?  A few
    hundred at best?  Is it worth forcing the other 119,000 employees
    to be knowledgable of physical network addresses so that we need
    only contend with name/node conflicts?
    
    Al

    First off, I have a gripe, not with VTX ELF per se, but with this
    notesfile.  Note 966 is titled "News about ELF V2", yet it reads
    more like propaganda about why we should be really happy to be
    getting less-than-expected functionality.  Since the note is write
    locked, there is no chance of rebuttal, so I fail to see why this
    is even in NOTES rather than in VTX, which was intended for things
    like announcements (or pronouncements as the case may be).
    
    Second, I would have to ask, how many people have bothered to inform
    their respective DIS organizations of the shortcomings of the new ELF?
    I'm not sure how it works worldwide, but in my area one of the DIS
    measurements is customer satisfaction.  I'm a customer, and I'm NOT
    satisfied with the idea of accepting lesser functionality, so I will
    make that plain when satisfaction survey time rolls around.
    
    Finally, while the arguments about the functional and technical
    deficiencies of the new ELF application have been noted at length,
    the real question that I have is:  Whose design and solution method-
    ology was used that ever allowed an application to be developed that
    had *less* functionality that its predecessor????  If I went to a
    customer and said, "Let me sell you this VAX to replace your XXXX,
    and oh, by the way, it does less than the XXXX did", I would be
    laughed out of town.
    
    Geoff
    

Re: .-1,

>    Note 966 is titled "News about ELF V2", yet it reads
>    more like propaganda about why we should be really happy to be
>    getting less-than-expected functionality.  Since the note is write
>    locked, there is no chance of rebuttal, 

What did you expect from the folks who are cramming VLF V2 down our collective
throat?  I presume you noticed that the author (and locker) of 966 is the
product manager for ELF V2...

        This moderator has a strong inclination to return topic 966
        to its author, but I'm waiting to get a "sense of the meeting"
        from my fellow moderators.  If you have concerns, please use
        MAIL to state them to all the moderators.  Please don't go
        down a "topic 966" rathole in this topic.
        
        Thanks
        
        Tom
        DIGITAL co-moderator

    RE: .139  "sense of the meeting"
    
    Tom, if you can tolerate a little non-sense from a non-moderator, I
    would suggest that, perhaps, the ELFen folk think the nodename IAMOK 
    should be pronounced as "I am OK" while the user community thinks it
    stands for "I, amok".  ;-)
    

    Well if it is under review shouldn't it be set hidden until
    further notice?

	I have, in my role as moderator, deleted topic 966 and returned
	it to the author. Questions regarding that action will be answered
	outside the conference.

			Alfred

I got into ELF V2 today and changed my password from the one that was supplied
by the ELF support folks.  I was not impressed to discover that after changing
my password, it would take 2 DAYS for the change to take effect!  I couldn't
believe it, so I tried to modify my entry.  I put in the new password and it
wouldn't take it.  I tried the old, and it let me in.

What kind of security is that?

Bob

    re .143
    
    Bob, you are obviously a pampered little brat who have never worked on
    an IBM VM ;-)

        [This note is being posted "anonymously" at the request of its  ]
        [author, who believes it is in his/her best career interest not ]
        [to be directly associated with the opinions stated here. I will]
        [forward any MAIL replies to the author if they are sent to me  ]
        [and clearly marked as a reply to this note.  Naturally, replies]
        [can be posted here; the author reads this conference.          ]
        
	I am writing this to head off/correct what could be a very bad
	situation. I strongly believe that if ELF V2 is released in
	its current form, it could have a negative impact on the ability
	of DIGITAL employees to GET THEIR JOBS DONE, and thereby a negative 
	impact on DIGITAL. 

	In addition, if ELF V2 is GIVEN to external customers as planned, 
	they are likely to perceive Digital's products as being poorly
	engineered. The fact that they wont be PAYING for it wont matter
	at all, since they will be given a solution that in itself is
	a *problem*. Customer perceptions ARE reality, and customers
	rely on reputation and word-of-mouth when they make their buying
	decisions. I believe that ELF V2 could damage DIGITAL's reputation.

	I have an intimate knowledge of the software development
	process at DIS, where ELF V2 was developed. There are some
	major differences between the way they develop software
	tools there and the way DIGITAL develops software products.

	DIS does not make software PRODUCTS, they make internal TOOLS,
	thus they are not held accountable to DIGITAL's many quality 
	standards. Their proposal, development, *approval* and final
	release process is internal to DIS and subject to DIS INTERNAL 
	standards *only*. The committee that makes ALL these decisions has 
	as a prominent member the DEVELOPMENT MANAGER for ELF V2.

	*************************************************************
	As a result, things that we take for granted in software development
	groups, like backwards compatibility, or a concern for how the
	customers feel about a product are IRRELEVANT in this context.
	*************************************************************
	They aren't SELLING ELF V2, so we aren't CUSTOMERS and it ISN'T a 
	DIGITAL product.
	*************************************************************

	You need to understand their process so you will understand
	how ELF V2 got the way it is... 

	THE *PROCESS AT DIS:
	--------------------
	o	DIS PAC (project approval committee) discusses need for
		a given idea for a network tool.

	o	DIS PAC assigns the development, if approved, to
		an engineering resource, usually DIS Digital Telecomm
		Engineering. Mark Endry manages this organization, 
		and is also a member of the PAC... I find it interesting that
		his name has not been mentioned in the ELF discussion, since
		he bears a great deal of responsibility for the look
		and feel of ELF V2.

	o	development begins.

	o	QA testing is conducted by DIS employees after development 
		has reached the right point. Formal, full time QA organization 
		does not exist as such, DIS employees are temporarily assigned 
		this task.

	o	a QA report is required by PAC, but the report is BY
		DEFINITION a RECOMMENDATION only. The final decision
		on whether or not to proceed is decided solely by the PAC,
		regardless of the conclusion drawn by the QA team.

	o	Field Test is administered by one or two DIS employees,
		at DIS sites, worldwide. 

	o	a FT report is required by PAC, but the report is BY
		DEFINITION a RECOMMENDATION only. The final decision
		on whether or not to proceed is decided solely by the PAC.

	o	if the PAC so decrees, the tool is distributed to
		DIS Sites on the easynet, and supported by Digital
		Telecom Operations at PKO.


	*reference: DIGITAL TELECOMM PROJECT PROCESS guide.


	WHAT DIS DOES NOT DO:
	---------------------
	YOU WILL NOTE that ALL of the checks and balances that occur in DIGITAL
	(MEM) software product development are absent here. A lot of us product
	developers might *think* that would be paradise, to be able to thumb
	our noses at DECstandards, SQM requirements, CSSE requirements, and
	CSC input into the development process, but clearly such a subjective
	process is prone to BAD judgement. It is human nature.

	(Frankly, SQM, CSSE, CSC, even HUMAN FACTORS groups' involvement can 
	seem like a real pain sometimes, but QUALITY and CUSTOMER SATISFACTION 
	are their concern. They dont let software development organizations
	have TOO Much control for this very reason. IN the end, we are grateful
	for their requirements, because we end up making something we can ALL
	be proud of, with no excuses required.)

	DIS does not use an OBJECTIVE, MULTI-ORGANIZATIONAL, QUALITY
	driven software development process, and the end result is
	tools with the look-and-feel of ELF V2.


	POSSIBLE CHANNEL FOR COMPLAINTS:
	--------------------------------

	There have been "petition" type group letters in the past
	week or two, to no avail. The individuals who have taken
	the role of spokesmen for DIS (Paul Devivo, John Beirne)
	are NOT empowered to alter the planned deployment. Their job
	is to make us feel somewhat better about a decision that
	has already been made.

	Since DIS does not make or sell products, we arent customers,
	and we clearly are NOT swaying them one bit with the noise
	we are making. The unfortunate truth is that senior DIS management
	may not be aware of most of our concerns; they are "shielded" by
	a layer of beaurocracy (some of whom sit on the PAC). 

	If you believe that ELF V2 will be detrimental to DIGITAL, or is
	worse than ELF V1, some might say that "do the right thing" means
	making DIS senior management aware of your concerns. I am neither
	encouraging nor discouraging this course of action, just making
	you aware of the situation and your options:


	*	The development manager for ELF V2 is MARK ENDRY @VRO

	*	Mark Endry's manager is PETER BROWN @VRO. He is also the
		manager of most (possibly all) of the PAC members.
	
	*	Peter Brown used to report to Bel Cross, but I believe 
		that has changed. Actually Ive heard that the VP that
		replaced Bel Cross in that function is a very reasonable 
		fellow. Maybe he is not being informed of these concerns? 
		(name unknown)

	NOTE: for the most part these individuals have disabled their VAXmail 
	access, as the result of an informal DIS policy regarding the use of 
	VAXmail and Nmail ( A wise choice. We all know how unreliable VAXmail
	is, and how awful Nmail is.  ;^)  Im kidding.). To reach them you will 
	have to use MTS addressing, of the form:

		MTS$::"VRO::First Last"

	where "First" is the first name and "Last" is the last name.
	Because MTS is store-and-forward mail, there is no guarantee
	how fast it will get there. Could be 5 minutes or 5 days, depending
	on the state of the intervening message router nodes. Heck, they
	may even get it more than once!  
         

    With corrective help from Alfred, here it is again:
    
    The following memo is being posted here WITH permission from John 
    Beirne.  John is the DT Network Applications Program Manager.
    
    My apologies for write-locking the original topic.  I was trying 
    to reserve space for future updates, and didn't realize my action 
    wasn't in the spirit of this conference.
    
    For those who have specific questions or feedback on ELF, please 
    use the ELF Notes Conference located at: 
    
    	IAMOK::ELF 
    
    (I never thought of pronouncing it I AMOK.)
    
    It is important that you utilize the ELF Conference since that 
    will be a starting point for future releases.  I don't regularly 
    monitor this conference, so if there is something you want to 
    bring to my attention, the fastest way is through the mail.
    
    Best regards, 
    
    Bob Igou 
    E-Mail Business Manager 
    DTN 273-5909 
    
    P.S.  This note is NOT write-locked. 
    
                   I N T E R O F F I C E   M E M O R A N D U M

                     V  R  O     A L L - I N - 1   S Y S T E M

                                        Date:      2-Nov-1989 02:04pm EST
                                        From:      John Beirne @VRO 
                                                   BEIRNE.JOHN 
                                        Dept:      Digital Telecom
                                        Tel No:    273-5150

Subject: ELF V2 Implementation

As we move into the implementation phase of ELF Ver. 2, I would like to 
recap the current status of the product, its future direction and how we 
intend to address some of shortcomings and additional needs identified in 
the Version 2.0 product.  

First, we believe that we have built a good product, one which is robust 
enough to perform on Digital's network and provide a reliable service, 
which the current version of ELF can no longer do.  We would like to thank 
the project team, the many Field Test participants, and all others who 
participated for both their efforts and continuing support in getting this 
service to the production stage.  

At the same time, we realize that there is still additional work to be done 
to provide additional performance and improvements to the user interface.  
Also, in light of the Digital network moving rapidly from a single to a 
multi-protocol environment, we need to expand the capability of ELF V2 and 
other essential network services to be able to accommodate the needs of all 
of Digital.  Although the new version has many capabilities that the old 
one did not have, ELF V1 was in many ways an intuitive service.  ELF V2 
differs in the way users interface to it.  We are sensitive to the initial 
pain that this change will inflict on our community and we ask your 
understanding as we go through the transition. We are committed to 
improving the service as we progress.

Additional changes in data content come along with ELF V2. In the old 
version the user had the flexibility to change virtually all personal 
information displayed. In ELF V2 , Personnel has reserved some fields such 
as Surname, Given Name, Organization, etc. which may only be changed by 
going through your Personnel System representative.  This is another source 
of aggravation for many but is essential for Personnel in order to insure 
accuracy of their data through exercising control on its source.  Again, we 
are both aware and sensitive to the effect of these changes but hope that 
you understand the rationale behind them.  
 
We have heard the disappointment from some segments of the Digital 
population that ELF V2 does not duplicate the ELF V1 functionality.  All of 
us take justifiable pride in being Digital employees and when changes in 
technology and systems force alterations in the way we do work, we would 
hope it would be to make our work easier right from the start. This is 
inherent at Digital: we are "solution providers".  Frequently in the 
beginning, however, we seem to cause more problems than solutions. It is 
only after we have migrated to the new that we begin to understand and 
appreciate what we have gained in the process. 

We have now reached the point where we need to move forward and provide ELF 
V2 as the production directory service.  We have been through two major 
Field Tests, made several modifications and have reached the point where we 
need and have the confidence to implement V2 as the production service.  
The viability of the existing ELF V1 is very limited and time is working 
against it. Further delays will just jeopardize the availability of any 
service. 

We have planned and scheduled a Phase II process for ELF improvements to 
commence after the implementation.  This Phase II process will contain 
several short-term interim point releases which will address the problems 
which are currently identified as well as making a provision for any new 
problems which may be identified since going into production.   

ELF V2 will provide the basis for our directory service over the next 3 to 
4 years and along the way, ELF will incorporate X.500 as its underlying 
retrieval mechanism and also provide Engineering with critical information 
on performance and management of a worldwide directory service which will 
help them as they develop a robust X.500 product set.  This process, like 
all long journeys, starts with the first step.

The decision to implement at this time was not an easy one.  We all wish 
that that ELF V2 were the ultimate system and could provide all the diverse 
needs that Digital has on day one of implementation.  However, due to the 
complexity of all the requirements and the realities of time and economic 
constraints we face today, this is not possible.  Again, we do have a good 
product that will provide a robust service and plans to make it even 
better.  We will be maintaining the communication linkages established 
throughout the Field Test process and providing information on our 
progress.  We ask for your cooperation and understanding along the way.

    
Re:  Note 957.135
    
    Martin's comment:
    
    "-- A system manager pointed out that the employee-settable passwords
   are stored in the database in human-readable text.  While the
   file is not directly readable by a malicious individual, were
   one to gain access, it is likely that real system passwords
   could be compromised. (Note 91.101)."
    
    Before you all go off the deep end...  The ELF password *is* encrypted,
    and is NOT a security concern.
    
    I believe security recommends keeping your various password different,
    that is your nodeA password, nodeB password, and ELF password should
    all be different.  Threfore if someone uncovers one, the rest aren't
    compromised.  

    
    Regards,
    
    Bob
    

    This information is nothing new, those following the ELF conference
    have see it.
    
    From my experiences, some things that make ELF V2 "look" like ELF V1 
    include, but not limited to:
    
    Defining a symbol or using the A1 device for the ALL-IN-1 key mappings.
    SYMBOL:= "VTX/DEV=A1/ASCII ELF"
    
    Typing the symbol, return, gets you to an ELF prompt, when finished,
    Gold and . on the keyboard to exit.  Difference - 3 keystokes.
    
    For searches, you might like the effects of for finding FIRST LAST
    something like
    ELF>F/G=FIRST/S=LAST
    
    If it doesn't return what you want, the next search might be
    ELF>F LAST     You may want to give it a SHORTER timeout as
    			these searches are typically fast.
    If you don't see what you want, start taking letters off the
    last name (all you're typing in) and using the trailing wildcard *, 
    until you do get something you want.
    
    Is that exactly like V1, no.  It's not sorted, and it will return
    ALL the entries when it doesn't timeout.  ELF V1 didn't.
    
    There are a few things ELF V2 doesn't do, there are some that
    take more keystrokes and/or time, there is quite a bit that
    V2 does that V1 doesn't.
    
    But aside from 2 general topics, I have yet to see anything that could
    be done in V1 that couldn't be done, after a fashion, usually
    with a silmilar solution, in V2.  Those are just message fields and
    sorted output; ELF v2 gives more places to write, that are searchable,
    which may or may not be a step up from one message field, that may
    give data overrun errors.
    	_bill

   Not to trip merrily down another rat-hole, but I thought the general
recommendation was that software products NOT store their own passwords, but
rather rely on authentication and authorization services in the underlying
operating systems.  Now, DASS isn't available now, but I'd translate that to
granting access to specific node/username combinations, rather than asking
for a password AT ALL.  I assume you can trust VTX's information about the
identity of the remote user who is attempting a change, otherwise you wouldn't
know the password to try to match.

					-mjg

    I probably missed this somewhere in all this torrents. But can some one
    recap Why is ELF v2 made to depend on VTX?
    
    I mean, VTX is great for information distribution. But ELF is not
    information, it's an APPLICATION! I really fail to understand, can
    someone enlighten?
    
    HW

    Some of the reasons, in no particular order, might include
    
    Digital has quite a bit of the commercial VTX market.  By using
    the product, it might increase that lead.
    
    VTX is a way to get I/O out to remote users.  It already offers
    11 terminal types, and DECwindos support, with more on the way.
    Did ELF V2 write a user interface?  No.  It uses our standard
    servers of a standard product.
    
    It is a standard product.  Did you have to learn yet another keypad,
    or protocols, or install anything?  Depends if you knew VTX up
    front, but VTX should be pretty easy to learn, there is not
    a lot that can go wrong.  ELF in particular has most of the keystrokes
    and what they do adjusted for you at most times.  It's also considered
    to be network efficient.
    
    Internally, VTX is on most systems, certainly at most sites.  Changes
    made centrally are already set up to filter out to the field. That
    sort of leads to the entire (world-wide) ELF system configuation
    can be changed from one spot.  Ever not get in?  If you made it to
    the ELF system, odds are we can track that.  If too many from some
    site can't get in we can modify it, and users need never know.
    
    By the way, once in the ELF system, ever fail to connect?  VTX
    is extremely robust in how IT manages connections.  There is no
    ELF code doing transport control.
    
    And not the least of things, it was there 2 to 3 years ago when the
    project was being designed, and there has always been tremendous
    support from VTX development/marketing/CSSEproduct and program
    management etc. over the years.
    _bill

The choice of VTX and DDS as underlying technologies for ELF are reasonable
from a management viewpoint, as many of the network and information distribution
problems are handled "invisibly" to the ELF developers.

I would certainly be interested in a VTX interface to the entire phone
book, so I can "thumb" back and forth in the book.

However, as many of us have stated at length, VTX/DDS (as currently implemented)
are unsuitable for the primary problem of searching for a name that may
be misspelled.  Whether the "right" solution is to extend DDS to add
additional search arguments (such as a phonetic-based hash code), or
to extend ELF to handle name-based searches better is a matter for
implementors.  Also, VTX is not necessarily a suitable presentation
mechanism -- this has also been discussed, both here and in the ELF
notesfile.

I don't think anyone would be bothered by the desing of VTX/ELF/DDS if they
worked the way users expect them to work.  Betsy Comstock (of the Dec
Human Factors Group) discussed this at length.

I don't think anyone would be angry with the ELF developers if they had
worked with the user community (Dec employees) to fix what was broken
(unavailablity) and improve what already worked.  Few, if any, of us are
skillful enough to develop something like ELF in a vacuum.

Martin.

RE: .151


Most of the systems in DEC have VTX?  Think again.  most of the timesharing
systems probably do, but I'd venture a guess that those with workstations
don't have VTX installed.


I certainly don't.  When I need to use VTX (very rare) I log into one
of our timesharing systems and use it.  When I need to use ELF, I use
it from my workstation (which I do often)  


I'm at the point with my dual-RZ23 system, where I can't install any
new layered products, because I'm out of space.  I'd be willing to 
bet many of hte workstation users are in the same boat.  The V1 elf
product was 64 blocks.  I'll almost bet my next paycheck that VTX 
is bigger.  I'm not about to throw out something that I use daily
in order to use the new version of ELF on my system.  What i'll end
up doing, is to SET HOST to a timesharing system (which by the way
is already overloaded) and keep a VTX ELF window around for the times
that I need to use ELF.  A wonderful use of the resources on the 
timesharing system.


Another major drawback of the ELF V2 interface is the lack of a DCL
interface.  Right now, I have the foregin command FIND defined to 
look for someone.  FIND BILL JACKSON is much easier than VTX ELF,
wait for a menu, type FIND BILL JACKSON and then remember to use
PF1-Enter instead of return.  I'm sorry, but that's NOT an improvement


-bill

    And, at least on the system to which I am attached, if I don't wait
    for VTX to paint the entire screen before typing the f[ind] . .
    I get garbage and have to:
    1.	exit VTX
    2.	log off
    3.	turn my terminal off
    4.	turn my terminal back on
    5.	wait for it to warm up
    6.	log back in.
    
    I *HATE* to wait for a menu when I know the next command and VTX
    requires that.  ELF v. 1 didn't.

Re: .151:

Bill, do you really believe that stuff?  I'm sorry, but if you do, you're
seriously out of touch with what the customers of ELF need.

    
>    Digital has quite a bit of the commercial VTX market.  By using
>    the product, it might increase that lead.

Is the purpose of ELF V2 to market VTX, or to provide a valuable service
to the Digital user community?  Lets get our priorities straight here.

>    VTX is a way to get I/O out to remote users.  It already offers
>    11 terminal types, and DECwindos support, with more on the way.

How many workstations can afford to have VTX installed on them?  My 6 meg
VS2000 with 1 RD54 needs a shoehorn just to get the essential stuff on!

>    Did ELF V2 write a user interface?  No.  

That is obvious.  Maybe you should have, but that is not the main point.
Telling a user that a new product is better because "it was easier to code
that way" rarely holds up.
    
>    It is a standard product.  Did you have to learn yet another keypad,
>    or protocols, or install anything?  Depends if you knew VTX up
>    front, but VTX should be pretty easy to learn, there is not
>    a lot that can go wrong.  

Excuse me?  I have been using VTX (when necessary) for years.  Does that
mean that I know to use FIND/S=mumble/G=fratz/FUll to get the same information
that a simple FIND MUMBLE FRATZ used to give me?  Not likely.
    
>    Internally, VTX is on most systems, certainly at most sites.  

Maybe they are on most DIS systems, but one of the problems is that you
are taking a tool that was designed internally BY DIS, FOR DIS with input
(apparently) only FROM DIS, and making it a Digital-wide application, forcing
systems that have nothing to do with DIS to install it.

>    Ever not get in?  If you made it to
>    the ELF system, odds are we can track that.  If too many from some
>    site can't get in we can modify it, and users need never know.

Yes, I've not gotten in.  VTX is not always available.  From the user's
perspective, there is no difference between that and ELF being unavailable.
If I can't get to it, it is unavailable.  If you have bound ELF to a less
reliable mechanism, that is a problem.    

>    By the way, once in the ELF system, ever fail to connect?  VTX
>    is extremely robust in how IT manages connections.  There is no
>    ELF code doing transport control.
    
There again, if I can't get to a VTX server, then ELF is unavailable, and
the users will most certainly know.

>    And not the least of things, it was there 2 to 3 years ago when the
>    project was being designed, and there has always been tremendous
>    support from VTX development/marketing/CSSEproduct and program
>    management etc. over the years.

Marketing support?  The only marketing I've seen is "Here it is.  Use it.
There's no alternative any more, because we took it away."  If you mean
VTX marketing, I'd say they did a very good job with you.

Kevin

    Re: Note 957.154 by TYCOBB::D_HILL

>   And, at least on the system to which I am attached, if I don't wait
>   for VTX to paint the entire screen before typing the f[ind] . .
>   I get garbage and have to:
>   1.	exit VTX
>   2.	log off
>   3.	turn my terminal off
>   4.	turn my terminal back on
>   5.	wait for it to warm up
>   6.	log back in.

    Does control-w help? For me, it repaints the screen and removes
    the garbage.
    
    Rich

    I went to install VTX on my workstation so I could use ELF.  The old
    ELF was indeed a tiny image, less than 100 blocks.  To install the full
    VTX kit requires something on the order of 30,000 blocks.  It is not at
    all obvious which of the subsets is the one you need just to access
    existing VTX applications elsewhere.  The installation guide was
    useless in this regard.  I finally figured it out by reading the
    KITINSTAL.COM, and discovered that even so it wanted 10,000 blocks.
    Not only that, but I had to raise one of my SYSGEN parameters to a
    value far higher than I had been running quite happily with previously,
    and then had to reboot for it to take effect so I could FINALLY install
    this "new, improved" application.
    
    When I did so, I found it reset my screen colors and wouldn't put them
    back upon exit.  I found that I could NOT find myself using the name
    that everyone knows me by.  I found that I got timed out trying to
    decipher the various fields that I was allowed to modify, so that I
    could put back in some of the information ELF V1 had that V2 lost.
    And then I found that the fields were too small.  And I found I rapidly
    tired of "No next form field available".
    
    Someone suggested using /DEV=A1 so that RETURN would work instead of
    PF1-ENTER.  Well, yes, it does, but when it tells you to press PF1-.
    to exit, it took me quite a while to figure out that the only key
    that would work was the period key on the main keyboard, not the one
    on the numeric keypad.  ^Z was ignored.
    
    ELF V2 is a serious productivity and resource drain on the entire
    corporation.
    
    					Steve

    re: whoever said I must have never worked on an IBM VM machine.
    
    Yep, you're right, the last IBM machine I worked on was a 360/40 running
    OS/MFT.
    
    re: .145 - the anon. note
    
    It's worse than you think.  I think it was mentioned that ELF V2 would
    be submitted to the ASSETS program.  If this is true, we won't be
    giving it away, we will actually charge people money for this.
    Of course, we could get lucky...Maybe it won't be up to whatever standards
    are required by the ASSETS people and will be rejected.
    
    Bob

   > When I did so, I found it reset my screen colors and wouldn't put them
   > back upon exit. 
    
    Steve, small consolation to all the VTX ELF hassles, but here's a
    login.com command line that will retain your screen colors when
    entering VTX...

$ VTX :== VTX/DEVICE_TYPE=VT240/ASCII
    
    Mike

    I was told by my management that there is a special mail
    account for employee concerns like this one:
    
    	SOCIAL::INVOLVEMENT
    
    And that mail to that account would be routed to the executive 
    committee with minimal filtering. 
                                           
    
    
    

    There's a position statement re: V2 in the ELF file.

        Re: .151

>    Digital has quite a bit of the commercial VTX market.  By using
>    the product, it might increase that lead.

        I can think of two effects of VTX ELF:

         -  Showing it to employees will make VTX look bad, and thereby
            discourage employees from ever considering VTX for
            anything.  This effect is already happening.

         -  Showing it to customers will make Digital and Digital's VTX
            product look bad.  Selling VTX ELF will damage Digital's
            reputation.

        					B.J.

    When is enough enough?
    
    	When the horse is dead.
    
    
    		This note now has over 160 replies, most of which
    		are not all that pleased with the new service.
    
    		The missing point is that this is a lot like peeing
    		your pants in a dark suit. It may give you a warm
    		comfortable feeling, but no one notices.
    
    		From reading the replies in this conference, it seems,
    		that VTX ELF was build without much contact with the
    		people that use it. It will be pushed to the hilt by
    		the people that have a vested interest in seeing it
    		promoted as far as possible within and outside the
    		company. And, another 160 notes in this conference
    		won't change that.
    
    		So, what can be done on a positive note, and who
    		is going to do it? If the answer is nothing and
    		no one, what's a new top?

Re: .163

I'll ditto that... ELF V2 will trash the good name of our VTX. I enjoy
using some of our VTX-based applications such as LOS (Literature Order System)
and Sales Update. These applications are a good use of VTX capabiliies, in
my opinion. ELF V2 may make use of the network and server functionality of VTX
but is a real dud when it comes to the user interface.

Lee

    re: space needed for VTX...
    
    I've been using a VTX image built back on a V4.7 system (no PAK
    required). Extracted the .CLD file from the kit, applied it to
    DCLTABLES, copied the exe to sys$system. Total disk space required
    about 650 blocks.
    
    I too think that there ought to be a "mini-kit" available.
    
    Dave

>    I too think that there ought to be a "mini-kit" available.
    
    The "mini-kit" is called ELF_FIND.  It consists of a 1-block .COM file
    and a 27-block .EXE file.  It's called from DCL, and doesn't require
    VTX to be installed on your system, though you need to know where the
    nearest VTX server is located.
    
    The ELF_FIND topic in IAMOK::ELF is topic 56.  If you're not a system
    manager you should read John Covert's reply which describes how to
    install ELF_FIND without privileges.
    
    ELF_FIND is not part of ELF V2.  IMHO it should become a supported part
    of ELF V2+, and enhanced to do more than just FIND.  For most uses of
    ELF, ELF_FIND probably will be the interface of choice for people who
    are comfortable with DCL command lines.  The problem with ELF_FIND
    right now is that the VTX interface has to be used to change the data.
    
    --Simon
    

    I have tried every combination of trying to find "Jerry Leichter"
    in ELF V2. It cannot find him. ELF V1 finds him.

    ELF V1 results:
    
    Name:  JERROLD S LEICHTER

Nickname:  JERRY

Loc/MS: DSG1-2/E6 Phone:
    235-8437  E-Mail: VIDEO::LEICHTERJ
    
    ELF V2 failed on the following combinations:
    
    1- Jerry Leichter
    2- Jerrold Leichter
    3- /c=Jerry /s=Leichter 
    4- /g=Jerrold /s=Leichter
    5- /c=Jerry /s=Leichter
    6- Leichter (This one worked it produced a different Leichter, WARREN
       LEICHTER) 

    Plus a few others...

    I would like to thank those who have worked so long and hard to bring
    forth a tool that doesn't work. 

Unless I'm mistaken, Jerry is no longer employed by Digital.

ELF V1 still shows him, since personnel updates aren't happening.

This problem will be solved in another 4-5 weeks, when ELF V1 is shut down.

/john

From HANNAH::LEICHTERJ Sat Nov 11 06:37:15 1989
Note-File: Marketing
Note-ID: 1031.14 (14 replies)
From: HANNAH::LEICHTERJ (Jerry Leichter)
Topic: Bugs in Intel 486 - and DEC's Image
Subject: (none)
Date: Sat, 11 Nov 89 06:37:15 EST
Lines: 28

[Text of note deleted.]

    That, or his Enet account is like his Yale account, and will last for
    ever.
    
    
    
    For those that don't know Jerry. He posts a number of helpful "answers" to
    the usenet infovax newsgroup (et al) from his account at Yale, where he
    has being doing postgrad work. 
    
    q
    

Therefore, he is employed and becoming Educated }8-)}.

				Bill

  I haven't read all the replies on this topic, so someone may have already
mentioned this:
  I have an anonymous memo titled "ELF V2 points of interest". It appears to
have been written for data center managers. It says, among other things, 
"[The] source code for ELF [V1] was lost some years back." Does anyone reading
this know why the source code was lost, and if this has anything to do with
the difficulty of maintaining ELF V1? 
  Another thing: I tried reading the IAMOK::ELF conference a month ago and it
ran extremely slow. I tried it at 11 pm and on a weekend, same problem. Does
everyone have this problem? 

	Last year when I was the Toolmiesterin I needed to verify that
	a person who submitted a tool to the ToolShed was a "real"
	employee of DEC - ELF V1 was the best way since I got tools
	from all over.  Sometimes I would only have a last name and
	a node name so I would use FINDNODE to locate the node and
	then check with ELF for someone living at the facility with
	the name given.  A few times there were people who stilled
	worked at DEC who had disappeared from ELF - That I could
	deal with by taking a second step of verification of
	employment.  This second step I only needed to take twice
	in the year and half that I took care of the ToolShed.

	I also used ELF to located old "owners" of old tools who
	had moved to new locations but had not notified the ToolShed.
	I was also able to find out that Jerry was not available
	since he was going to be at Yale but I could do x to reach
	him.  There were a number of other tool owners who I traced
	through the messages in ELF.

	I could go on and on about how I have used and relied on
	ELF to get me information about other employees from all over
	the world.

	I have not used the new ELF, I have a workstation and I
	have not/do not intend to install ELF V2 on it (I will most
	likely use the tool out of the ToolShed instead).

	BTW:  Last Spring (I think, it may have been the previous
	Fall) IDECUS I talked to the ELF V2 people about what they
	were DEMOing.  I told them in person and in writing that
	the two things that needed to be improved in ELF V1 was
	to the update process of employee change information and to 
	make it	easier to find someone with out having the correct spelling.
	I also stated that availablility was in issue with some people
	but I used a tool out of the ToolShed that fixed that for me.
	They told me, that both of these things were high priorities
	for ELF V2 but that they had not been able to attain them
	as of that time but that they would before ELF V2 was released.

	_peggy

    I wasn't going to complain about this because although ELF was really
    nice, and new ELF may not be, I don't use it much, but...
    
    To whosoever decided that "No ELF servers available" was a reason
    to replace it, I suggest the current service unavailable and page
    inaccessible messages more than make up for it.
    

    
    
    
    
    
    
    1 and 2 are different products. They also use different data bases! 
    
    
    Using ELF 1, I looked up a former DEC employee who left DEC six
    months ago.  He's not in the data base.
    
    Using ELF 2, I looked up the same person... Lo and behold he's
    still listed!
    
    What a feature.  Security will love it.
    
    d

    
    Another feature us 'Irish' guys like is the ability to spell O'Grady
    properly and improperly.  It seems that digital decided many moons ago
    that I was OGRADY, not O'Grady.  That's computers :-).  Well ELF V1 let
    me be O'Grady or OGRADY.  It even lets me be O;Grady, O"Grady, O-Grady,
    etc.  It is flexible.
    
    ELF V2 says I'm OGRADY.  Geez, what progess, huh?
    
    gog :-(
    

.......

    Last interactive login on Tuesday, 21-NOV-1989 13:16

.......

Ripper_> elf
That information is not available.
Ripper_>
Ripper_> lo
  AMES         logged out at 21-NOV-1989 13:17:31.80


(Australian EDT, lots of VTX control charaters removed)

    Re: .180
    
    If you are looking at the VTX server on SNOV00, VTX ELF should now
    work (or at least start...)

    
    Howdy!
    
    I'm a technical midget, and like it that way.  I called my System
    Manager to complain about the new elf, and he directed me here.
    
    VTX Elf is so unfriendly, I've resorted back to using my phonebook.
    
    My specific issues are:
    	-  have to go through a screen display in order to initiate a
    		command (i liked "elf find person's name")
    	-  can't type ahead; the screen locks, and you have to ctrl c
    		out of it
    	-  the PF* key plus another key to initiate a command drives
    		me nuts; as a touch-typist, it's awkward to stop to
    		hunt and peck a key combination for a simple <cr> action
    	-  people's node names seem to have disappeared, no longer 
    		included in part of the display
    	-  too stringent around spelling; now I have to know the first
    		name exactly in order to access an elf entry.
    	-  it's slow
    
    So, bottomline, I want my old ELF back!!!  I miss the little guy.
    
    Susan Olski
    Colorado Springs
    

Your system manager should be (or become) aware that there is a DCL-based 
interface to the new (but definitely not improved) ELF in the toolshed, called
ELF_FIND. 

To discuss the new ELF directly with its developers (or at least with its
defenders), try the conference IAMOK::ELF (pronounced "I [run] amok", rather 
than "I am ok").

Try the 'other' OS .......  34 seconds for the answer delivered around the
world.  VTX (if it started at all) would take that long for the first screen to
appear!

Transcript follows;

Richard.
-------------------------------------------------------------------------------

Ripper_> elf


%VTX-E-BRKNOLI, VTX login has not succeeded

Ripper_> set host redbck


Ultrix-32 V3.1 (Rev. 9) (redbck)

login: ames

Password:


Ultrix-32 V3.1 (Rev. 9) System #13: Fri Oct 27 04:27:27 EST 1989


                      Unauthorized access prohibited

redbck> elf richard ames

Trying ELF gateway on CIRCUS...connected (delay=0:07)

Response from CIRCUS (delay=0:27) ...

               Name:  RICHARD AMES
          Telephone:  +61 2 965 7747
        DECNET mail:  GIDDAY::AMES
   Intrnl Mail Addr:  STL-3/5
           Org Unit:  CSSE - South Pacific Region, Sydney - Australia
           Position:  Manager

redbck> 

REM-S-END, control returned to node _RIPPER::
Ripper_> log
  AMES         logged out at 29-DEC-1989 12:33:42.79

    Re: Note 957.183 by STAR::BECK

>Your system manager should be (or become) aware that there is a DCL-based 
>interface to the new (but definitely not improved) ELF in the toolshed, called
>ELF_FIND. 

    Note that you needn't wait for your system manager to install this
    tool. You can install it in your own account, as well, which I have
    done. It makes V2 ELF seem much more like V1 ELF. 
    
    However, even this tool does not overcome the one big failing of V2
    ELF, which is that you *must* spell the name correctly. However, you
    can use a wildcard in the name (e.g. Nels* to find Nelsen and Nelson)
    to help a bit with this. 
                             

    re: .184 by GIDDAY::AMES
>Try the 'other' OS .......  34 seconds for the answer delivered around the
>world.  VTX (if it started at all) would take that long for the first screen to
>appear!
    
    In fact, there is still only one distributed interface to ELF v2;
    that's through VTX, and even going around the world you are still
    using the same VTX service down the line as everyone else.
    
    It's not "usual" for VTX to need 34 seconds to display a screen,
    your support staff may want to look into anything over 10 seconds
    at worse.
    _bill

All these DCL-like interfaces are great, but they don't help the poor people
trapped in the DIS ALL-IN-Prison.  They are just plain stuck.

Bob

    ELF was there most of the time.  VTX ELF has been "that service isn't
    available" for a couple of days (for me, anyway).  I've reverted to my
    DEC (paper) phone book for the first time in years.  Sigh.

Has anyone been able to use the new and improved ELF in the past week?
I too have only been getting "That service is not available".  It took
a half hour to dig out my phone book.

John

    re .187
    Check the IAMOK::ELF conference.  I think there's reference of
    a method of linking the ELF_FIND to ALL-IN-1.  We have it running here
    and it works fine...right from my mail menu.

There again, you're assuming the ability to change something <ANYTHING>
on one of these systems.  The locked-down systems being discussed are
rigorously locked down.  It would involve convincing someone in DIS
to figure out how to do it, and to actually get around to it.
As I write this on a DIS machine running VMS 4.7 (with NO eta for 5.x),
the above doesn't seem very likely.

Kevin

    Who is in charge of VTX ELF data?  VTX ELF invented a DTN phone number
    for me, but we do not have DTN here.  This is costing the company money,
    because other people are wasting their company time and paying for 
    unnecessary long distance phone bills to call me with the wrong number.  
    I tried to correct this, but the most that I can do is to add the correct 
    non-DTN phone number to VTX ELF.  The DTN number is bolded, which means 
    the "data is provided by Personnel and may be changed through your 
    Personal Services Administrator", according to VTX ELF.  I talked to my 
    PSA, but she could not find the incorrect DTN information in my personnel 
    file.  I am now waiting for my local VTX ELF administrator to see if this 
    can be fixed.
    
    I have also found that VTX ELF is not available more than half of the
    time I tried since the beginning of this calendar year.  I hope this is not
    intended to be THE solution to stop my incorrect DTN phone number from
    being used.
    
    Charles

The place to report this is _not_ in this conference, but in the IAMOK::ELF
conference.

When you report it, you might point out that a scan of people at UCS reveals
293, 408-864, 521, 415-853, 273, 291, and 223 numbers (all wrong) along with
a few 415-691 numbers (correct).

In the ELF conference you will learn that this data comes from personnel for
non-US employees and from EMAS Special Services for US employees.

The way to update this field is to submit the extension change form in the
corporate directory, or to send mail to ICS::DIRECTORY.

The reason that it appears in the DTN field is that the corporate directory
uses a single field, and provides ten-digit numbers for people at locations
not on DTN but with direct-inward-dialing.  If the ten digits were actually
there, it wouldn't really matter that it says DTN, because people at U.S.
DTN sites still would dial the same way (8+) and people at European sites
would recognize that the number is not really DTN and figure out how to
call it.

By the way, this same incorrect number appears on your entry in the phone
book.  For all ELF V2's shortcomings, one thing that it does provide is a
way for you to become aware that there is something wrong with your entry
in the corporate databases.  Did you previously work in Boxboro, and was
this your number there?

/john

    FLAME ON!!
    
    I have used ELF in place of a phone book and it has worked quite well.
    In fact I do not have a DEC phone book in my cubicle.
    WHEN I TRY TO ACCESS VGTX ELF 70 PERCENT OF THE TIME I GET THE MESSAGE
    "THIS SERVICE IS NOT AVAILABLE"  THAT STINKS!
    WHEN ONE SERVICE IS REPLACED BY ANOTHER IT SHOULD BE JUST AS
    AVAILABLE!
    
    SET FLAME TO SIMMER:
    
    NOW, I do not know if this is because the VTX server at my facility
    is not working, or if it is somewhere down the line, but I do NOT 
    think it should matter.  
    
    AN UNHAPPY USER,
    
    	FRANK 

    RE: .191
    
    For the record, I just contacted the folks who run RIPPLE (a USIS
    office (A1) system). They will be upgrading to VMS V5.1 by the end of
    the month. I'm sure they would be more than happy to talk to you about
    adding the ELF_FIND stuff to A1 (or whatever you are looking for). Just
    give them a call. If you need a name, send me mail.
    
    -Glenn-

re: .193

Thanks for the pointer, John.  I have checked IAMOK::ELF, and have suggested 
that ELF explains that DTN numbers can only be changed by chaning the 
corporate phone directory first.  I also passed along the information
on the wrong UCS phones.

Yes, I did work in Boxboro, and my current (incorrect) DTN number was my
old number.

Charles

    Put me down as another unhappy user of VTX ELF.   It's so slow as
    to be useless.   It has bizarre user interface features like right
    arrow to go DOWN a field, return key which doesn't do the same action
    as 99% of the software in the world, and overstrike instead of insert.

    Worse, I can't find people, who I know are real, using their name
    exactly as it is in the phone book and on their badges.   Why did
    we ever do this to ourselves??????????

You would think that it would be at least as clever as the
old ELF.

Bah Humbug !

RE: .195 RE: .191
                          -< Have you tried asking? >-

For the record, I have asked, and I got the same answer you did:
"We will upgrade within the month". I have been getting the same answer
since July. My comments stand.

Kevin

    Register me as not happy with the VTX ELF.

    Oh, come on!
    
    This new thing is good for you! You're all just "resisting change".
    "They" know what's best for you ... trust "them".
    
    
    (GIMME BACK THE OLD ELF BEFORE I KILL SOMETHING!!!!! ! [I've already
    kicked the dog screamed at the kids and beat the wife])
    

       RE: .197
       
>Why did we ever do this to ourselves??????????
       
       	Correction please.. "Why did DIS do this to us??" is the better
       question.
       
       							mike
       

    Here is the login message that is now appearing on my cluster.
    
    4-JAN A new ELF system has been installed corporate-wide,
          to use it type "ELF FIND/FULL first last". Please
          mail any problem reports to IAMOK::YOST.
    
    I tried this out this morning. The first time I tried, I simply
    got a 'network connection rejected' message. I re-issued the
    command and, after a short delay, received the full listing for
    the individual leaving me at the $ prompt. This worked super,
    but I'm sure it has something to do with how your local software
    support folks implement the 'features'.
    
    Bob Mc

    
    
    Could somebody please tell me why ELF has been made more difficult use?
    It seems that we have again made something less user-friendly to use in
    the name of improvement.
    
    Bob

re: .204:                -< ELF Why more difficult? >-

My understanding is that the ELF V2 developers decided to use off-the-shelf
software for as much of ELF as possible:

-- VTX for the user interface and overall managment of the process.
   This means they don't have to figure out how to send escape sequences
   to the various terminals, and the client-server software is widely
   distributed (to VMS systems).

-- DNS (?) for the database.  (This significantly simplifies the problem
   of distributing and managing the database.)

Unfortunately:

-- VTX has, for people accustomed to command line interfaces, a bizarre and
   counter-intuitive human interface.  According to the original VTX architect,
   it was never designed for this type of task.  It is also unavailable on
   Ultrix.

-- DNS does not know how to search for "almost-matches" (the sort of thing
   that any software dealing with human language needs).  Also, it has a
   strange algorithm for searching for multiple matches that make it so
   slow as to be unusable in many situations.

The bright side:

-- An Ultrix client is available (a hack of unimaginable proportions).

-- The ELF folk are aware of the problem (understatement) and, to judge
   from the conversations I've had with them, plan to fix the search
   problem.  (I have no idea when.  There are a lot of algorithms out there,
   and a real need for human factors testing and consultation: I don't know
   what they're planning).

I would hope that the developers will continue to be open to criticism
from the user community, and will use "engineering standards" such as
the phase review process when implementing ELV V2.1.

Martin.

    At one time, I heard talk of productizing ELF like we have done or plan
    to do with other internal tools.  I sure hope that we don't productize
    V2.  I too have started using the phone book because:
    
    o I can shuffle pages faster.
    o I don't have to terminate command lines with two strokes.
    o I don't have to know exact spelling.
    o plus, most of the other reasons mentioned in this topic.
    
    It would be criminal to ask customers to pay for this.  DIS should be
    paying us to use it.

The underlying directory service which is a major part of ELF's problem is
DDS, not DNS.

/john

Well besides using my phone book as the rule these days, I also get mail
from a number of people with old phone books asking me if I know the phone
numbers for other people because the former can't find the latter using ELF
V2. This is absolutely crazy! 

    
    ELF V2 must be costing this company MILLIONS in miscommunication!
    
    The folk on "IAMOK::" may be ok, but the rest of us are not!
    

    RE .209

>    The folk on "IAMOK::" may be ok, but the rest of us are not!

    Typical misunderstanding. It's not "I Am OK" - it's "I [run] Amok".

    Ho Ho - except this ain't funny any more.

    Alternate interpretation of the new ELF:   Force everyone back to the
    phone book, using voice communications on DTN instead of electronic
    MAIL (and Notes, or whatever) on the EasyNet.  That makes the bandwidth
    available for the necessary functions of monitoring, security probes, 
    etc.
    
    I've found there is alot of new information in the DEC phone book since
    I seriously used it last.
    
    	An alphabetic list of facility codes
    	A list of DTNs by prefix to facility
    
    The only thing the DEC phone book is missing is node address.
    
    The new ELF won't find half of the people I want to find.  Once I go to
    the phone book to find that Al's first real name is Albert, ELF didn't
    have a node for him.  Missed a few Ed's, and others too.
    
    Oh well, back to analog communications.   I'm not the only one.  I've
    noticed I get *lots* more phone calls than I did a few months ago.

><<< Note 957.212 by ULTRA::PRIBORSKY "All things considered, I'd rather be rafting." >>>
    
>    I've found there is alot of new information in the DEC phone book since
>    I seriously used it last.
    
>    	An alphabetic list of facility codes
>    	A list of DTNs by prefix to facility
FYI: These are on line in the DTN conference, note 2.2
     (on AYOV18::DTN)

>    The only thing the DEC phone book is missing is node address.
Also on line. You can get this list in:

ANCHOR""::NET$NODES:MININODE.LST

Beware, it's 5000+ blocks though!

	Craig

	RE: .213 Yes MININODES.LIS is on line but I think what Tony was
	refering to as missing in the phone book was the spicific node
	name for spicific people. That isn't in MININODES.LIS.

			Alfred

        Geeze, I just love being at home and trying to find out a node
     address.  I had to deal with VTX ELF last night and I was ready to
     strangle the computer!  I got tossed into some sort of bizarre
     character set that is totally illegible.  New feature, perhaps?
     Great, reset the terminal and try again.

        I asked around at work today and others have encountered this
     wonderful phenomenon, too!  They were as unimpressed as I was.

     	Vicki

    And you think you've got problems. ELF used to give me reference
    to all DEC people - US,Europe,GIA,UK. 
    Now I just get the UK set. Trouble is I spend most of my time
    nattering to US and Europe ...
    
    To be fair, I believe V2 solves a lot of ELF-internal problems.
    I'm sure it's vastly more efficient at what it does - whatever
    that is. Perhaps someone could let us in on the secret.
    
    Mike Day
    

    It's very efficient - at returning us to the 1950's.

    VTX is unavailable today.  Since this morning at least.
    
    Bashing ELF V2 seems in vogue.  (I don't like it either.)  And a LOT of
    negative stuff has been said by a LOT of people.  
    
    Is there anyone who defends VTX ELF???  
    
    tony	(just curious)

>Now I just get the UK set.

You are the first person to report this problem.  Be sure you're right, and
then post this in IAMOK::ELF.

/john

    Well, I'll confirm that the GIA people have disappeared off the version
    of ELF I have access to, and I'll add my voice to the general opinion.
    
    BOOOGGGUUUSSSSS!!!!!!
    
    Susan

All countries?  I can still find people in Hong Kong, Japan, Australia,
New Zealand, and Canada...

/john

    re .221
    
    Thank God!  I was beginning to wonder if we were really here, or
    whether we were just fading figments of an I-run-AMOK imagination!

    Re .216 .219 etc ..
    
    FIND W XYZnnnn  works. (This is using ELF_FIND.EXE) .
    Anyone can be found (GIA UK US EUR). As long as there is a VTX_SERVER
    
    However, the "partial key search" VTX XYZ , only throws up (if you
    will excuse the apt phraseology) , XYZs that are in the UK ...
    
    I'm sure there is a logic pattern somewhere ....
    
    Not posted in IAMOK. I don't want to be accused of increasing stress
    levels any further.
    
    Mike Day
    

    Some of the problems locating people with ELF V2 may be caused by
    incorrect use of the FIND command.  ELF V1 used the command syntax:
    
    		FIND user@loc
    
    In ELF V2 the syntax is:
    
    		FIND user/LOC=loc
    
    If the "@" syntax is used in ELF V2 an "ELF entry not found" error
    message is displayed.  I found this to be very confusing, especially
    since I had found the same person in ELF the day before!

         <<< Note 957.224 by WELKIN::ADOERFER "Hi-yo, Server! Away!" >>>
>    Don't worry too much about the stress.  Very few elf support types
>    are reading this string since it became the Andy L show.  By the
>    way Andy - *yawn*.  :-)

    I beg your pardon? I haven't sought to make this topic my "show" - and
    hadn't indeed entered anything in this topic from 4-Nov to 13-Jan.
    
    If you think that 14 replies out of 200+ makes this my "show" then your
    math grades must have been terrible.
    
    If I'm unimpressed with the ELF, I'm *extremely* unimpressed with the
    attitude that you display with such silly comments. They aren't funny.
    
    - Andy

    Several sites/servers have their own telephone directory services
    maintained by people at those locations.  Usually, they only
    have local data.  ELF_FIND would bypass all those, or at least
    the ones known so far, and try to go to ELF V2.  (by the way,
    you can not only define vtx$server, but vtx$server02 through
    vtx$server09, so if your net is up, you should have 9 chances).
    
    I'm not aware of a "partial key search" that can possibly use
    a different set of data, so I suspect you got into some other
    application.  You could post a sample of the screen you get, but
    of course, the proper place would be the IAMOK::ELF notesfile,
    or your local ELF admin., or your vtx support staff, or possibly
    your system manager or area network manager(s).
    _bill

    thanks to the new improved ELF, the teleophone book and rolex have
    been put back on the desk.  Another leap forward for man kind.
    
    I try to use the new one.  Funney thing happen the other day.  I
    had three people to look up.  I started using ELF.  Whe it appeared
    to be operating so slow, I was able to look up all three in the
    phone book before I could move from the main menu to the find menu.
    And I'm not running off a PDP-8S!
    
    And boy, its not only bad for the poor speller [like me], but for
    those who didn't realize they had to register all versions of their
    name, it becomes impossible to find them. i.e. doug, douglas  I
    would have never thought of having to use douglas for the person
    I only know of as Doug!
    
    I want to personally thank those who created and forced this
    improvementdown my throat.  My your future salary increases be
    reflect the effectiveness of |ELF V2, and may I not be the reviewer.
    
    ed
    
    

>    I want to personally thank those who created and forced this
>    improvementdown my throat.  My your future salary increases be
>    reflect the effectiveness of |ELF V2, and may I not be the reviewer.
    
    May they one day be forced to ask you for a job...

    There are a couple of simple things one can do to make the new elf a
    little more productive: 1) use the * as much as possible, and 2) use
    the /full modifier.
    
    For example, to look up Doug Blaise (made up) and get as much info as
    possible about him you would type:
    
    f/full doug* bla*
    
    if this times out then try:
    
    f/full doug* blai*
    
    This finds Mr. Blaise when his first name is entered as Doug, Dougy,
    Douglas, etc.  Of course, if you thought he spelled his name Blase,
    then you are SOL if the first case times out.  Once I started using
    this procedure (*) I didn't find the new elf any worse than the old elf
    in locating names and it certainly is available more often.  What I 
    really don't like is having to hit PF1 enter.
    

To get around some of the limitations of the new Elf, I wrote a front-end
to ELF_FIND that does a few things that I wanted:

-- sorted output
-- no escape sequence stuff
-- output can be written to a file for post-processing (Unix > redirecton)
-- if I can get a database, I'll try to do some sort of phonemic search
   (which is what I'm really interested in).

It's still very much "under development" and changes more-or less hourly.
Bug reports (and, especially, bug fixes) are always welcome.  The readme.txt
is attached.

Martin.

Announcing Elph 0.6.  Not quite everything you want in an Elf.

Sources:

	may14::pubroot:[elph.distribution]*.*
	All you really need are elph.exe and elph.mem

Prerequisite

	ELF_FIND (version 1.4) from the Software Toolshed.

More information, if any.

	MAY14::ELPH.NOTE

Limitations:

	Timeout is whatever the network gives us.
	Someone who understands Decnet might try making it bulletproof.
	Elph doesn't run on Ultrix (use Brian Reid's).  Very easy to do.
	Doesn't understand terminal width/height.
	Performance.

Improvements in this release:
0.5,.6:	Various bug fixes.
0.4:	Compiled to use a local copy of ELF_FIND.  This compilation
	does not need (or care about) the Ultrix server.
0.3:	Printout fixed.
	First name heuristic (see below).
	Can read from standard input.
	Can control server search list by setting ELPH$SERVER logical
	Better control of debug verbosity.
	Name changed to avoid conflict with an existing alf.

Advantages over ELF_FIND

	Comes with source code.  Has somewhat better search and display
	heuristics, will get phonemic matching when I create or
	obtain a database.  Output is sorted (assuming the actual
	data is correct). No PF1.

Disadvantages over ELF_FIND

	A midnight hack, so the only support is what you get from me or
	work out for yourself.

	Will disappear if/when the network police decide it's a security risk.

	Slower as it has to go through the Ultrix elf server (if compiled
	to go through the server, of course).

First name heuristic:

	"$ elph k* olsen" reads all "olsen" and searches for the match
	internally.  This works around a problem in normal Elf (which
	times out if you try this).

To Do:

	Local database.
	Decompile ELF_FIND to connect directly to the server.

Limitations:

	You must have ELF_FIND up and running, see elph.mem for the
	gory details.

	Sometimes the interprocessor buffering gets messed up.  Probably
	a bug in my handling the mailbox stuff.  Help would be appreciated.

	Error messages and progress indication could be better, much better.

Your suggestions are welcome.  Additions, too.

Installation:

	1. copy elph.exe to a known disk::[directory].
	2. make sure ELF_FIND is installed (try it!)
	3. make sure your login.com defines the ELF_FIND parameter:
	    $ define elf_find$par "FTFF"
	4. Define the elph foreign command or add it to your login.com
	    $ elph :== $disk:[directory]elph.exe
	5. Make sure you have enough privilege to start a subprocess.

Options and other information in elph.mem

Martin.

Well thanks to ELF V2, I am finally using the personal names in Notes
and MAIL for what they were intended for.

    I initiated an SPR on ELF FIND last week. It, apparently, quits 
    with the first match that it finds.
    
    There are 2 Robert J. McClure's in DEC and to make it worse,
    we are both in SIMG. If you type "elf find/full bob mcclure"
    you will get the one in ZKO, only. If you type "elf find/full
    robert mcclure", you will get me only. The 'search given name'
    field for the guy in ZKO lists "bob, robert, robert j" in that
    order. Mine lists "robert, robert j, bob" in that order. The search
    kicks out on the first exact match. You will not get both listings
    unless you go through VTX ELF.
    
    

>    I initiated an SPR on ELF FIND last week. It, apparently, quits 
>    with the first match that it finds.
        
        No, it doesn't stop at the first.  It's just how your employee
        record is set up in the corporate database.
        
        Doing a FIND/FU BOB MCCLURE makes sense because this is the
        usuable way to use a utility like this.  Unfortunately, that's
        not the way it was decided to implement it.  The above command
        is equivalent to FIND/FU/COMMON="BOB MCCLURE".
        
        Unfortunately, the only way to get a decent search seems to be
        to do FIND/FU/SUR=MCCLURE/GIVEN=BOB.
        
        Look at the help in ELF for /SUR, /GIVEN and /COMMON.
        
        sigh . . .
        
        _Mike

You haven't found a new problem.  The problem you ran into is the subject of
TOPIC FOUR in the IAMOK::ELF conference.  A well known problem.

/john

The "1.0" release of elph runs on either VMS or Ultrix.  (But Ultrix
users may be better served by Brian Reid's elf.)  This version does
not "connect" with ELF_FIND, but talks to VTX directly.  Unfortunately,
I figured out what to send by dumping I/O buffers and experimenting, so
there are lots of magic numbers and probably a few bugs.

Advantages over ELF_FIND

	Comes with source code.  Has somewhat better search and display
	heuristics, will get phonemic matching when I create or
	obtain a database.  Output is sorted (assuming the actual
	data is correct). No PF1.  Has a simple terminal pager.

Disadvantages over ELF_FIND

	A midnight hack, so the only support is what you get from me or
	work out for yourself.

	Will disappear if/when the network police decide it's a security risk.

See MAY14::ELPH.NOTE for more details.

Martin.

re: .233,

	My ancesters were smarter.  They chose the less common spelling!  ;^)

				    -davo

	    (who still gets confused with David E. McClure in MR01)

re: .233:

One reason why MPBS::MCCLURE can't be found in Elf as "BOB MCCLURE" is
that his alternate given name was entered as "Bob".  Using ELF_FIND,

	$ ELF_FIND """Bob""" MCCLURE

If you get your PSA to re-enter your alternate given name as BOB, you
might get better results.

Martin.

    Oh, no!  Does the capitalization have to match, too?

re: .239:

    Oh, no!  Does the capitalization have to match, too?

I'm not completely sure.  If the entry is "FOO", both "FOO" and "foo"
seen to match, but if it's "Foo", only "Foo" matches.  It might not be
a bad idea for the thing that updates the database to force all upper-case.

One other point: I don't think that the database understands Multinational
and/or Latin-1 yet.  I'm sure that that change will be traumatic.

Martin.

       
       
       
       	Gee Bob, I had no problem looking up your number.....
       
       
       	In the phone book.
       
       							mike

    Sure Mike. Especially after I gave you a clue regarding Bob &
    Robert.
    
    What I really like, is when someone changes location and their
    ELF listing has the correct(new) location but their old telephone
    number. Then there's the guy that doesn't have an entry in the
    phone book and his ELF entry has the correct location but lists
    a bogus phone number. At least I was able to VAXMail him a message
    asking for his correct number. Typical DEC documentation. The
    info that you're looking for is only referenced by a "See Mumble-
    fratz Manual". Which, of course, is the one that happens to be
    missing from your set.
    
    Bob Mc

    re .233
    
    Gee, I would have looked for you with the following:
    
    f/full rob* mcclure
    
    And both of you show up.  As I mentioned previously, I use the *
    as much as possible and have little trouble in locating people with
    ELFv2.  Why use the difficult method when the much simpler one of 
    using an * is available?

    re .243
    
    Interesting, I would never think of trying that. After all, everyone
    knows how to spell robert. The point, I think, is that most roberts
    seem to go by bob. It certainly is more popular than rob, then there
    are the few that go by bert. I still think that *most* folks would
    try bob and then robert.
    

    re .244
    
    Interestingly, it works for bob* and robert*, must be a feature!
    
    However, using your logic I probably would have tried bo* not knowing
    it worked for bob*.  The reason I used rob* is I assumed the elf
    database tends to use given names (Robert) before nicknames.  If I were
    searching for a bill, I'd try wil* before bill*.  But that's just the
    way I'd do it.  The key so far seems to be to use the *.

    This all demonstrates a problem with how we do field testing here at
    Digital.  We measure field test by number of QARs answered and whether
    any have remained open; I don't know of any groups that measure field
    test by percentage or absolute number of responses in the "fixed in
    next release," restriction," and "user error" catagories.  We should
    probably be looking at these non-answer QAR responses, and if there are
    too many, seriously ask whether the product is suitable for release.
    
    	HRB

    
    
    For those interested, I wrote a little utility called DWELF based on
    ELF_FIND. It a little DECwindows prototype (with bugs..:-)).
    You need to have VTX running on your system to be able to run it.
    
    The utility can be found on VISA::SYS$KITS:[DECW$UTILS]DWELF.SAV.
    For more information refer to the notes conference DW_EXAMPLES (on
    ELKTRA) or on (IAMOK::ELF)
    
    Rgds.
    
    fg/

    After all of the complaints about the first VTX version of ELF I think
    the time has come to redress the balance. 
    
    I have just been using ELF V2.1 and find that it is very quick (I think
    quicker than ELF V1) and it finds mis-spelt names and partial names
    with few problems. Well done to the development team.
    
    Gary Harding

While I sympathize with complaints about the new ELF, I also have
painful memories about trying to cope with 'classic' ELF.  All in
all, the 'upgrade' from old to new has all the passion/excitement
of many previous IS endeavors.  It's going to be a case-study, no
doubt.  Personally, I think the new one is ok.  There were really
terrible problems with the old tool.

People who are really offended by ELF V2 owe it to themselves to copy
the award-winning ELF_FIND from the Toolshed, for its kindler/gentler
'human interface'.

Looks like new ELF has uncovered mail-handling problems that were
masked by classic ELF.  My painful memories, by the way, have to
do with changing my office location twice in a period of a couple
months, and having to spend several days each time chasing ELF
servers world-wide attempting to create consistent/accurate data
for myself & others, struggling with 'server-unavailable' problems
and 'you're_not_registered_so_you_can't_make_changes' nonsense, while
the mail-handlers wouldn't forward paper mail because classic ELF told
them my address hadn't changed.

I got this gem this morning about VTX ELF, I think it's great!

  ** Posted with the original authors permission. **

	Craig

Subj:	A new classic from Bob Stewart on VTX ELF

From:	ALLOUT::STEWART "01-Jun-1990 1052"  1-JUN-1990 12:22:19.83
To:	@STFPLUS
CC:	
Subj:	*&^%$&^$%# VTX ELF &^%)^%#$@&%@

Its times like this when I think maybe 67 is a fair market price for DEC stock.

How can a multibillion dollar company, a leader in the computer industry, the
greatest networking company the world has seen, a company with pretensions to
beat IBM at databases and transactions processing, a company that makes believe
it is going to make money by being great at systems integration, produce such a
flaming abomination?????

Every time any text book wants an example of a simple database, one that any
student can understand, they use a company personnel database. There are only a
few fields, it does not change all that fast, there is no complex hierarchy, no
circular dependencies, etc. The things that have to be done to it are not
complicated, just add new bodies, remove old ones, update fields when someone
changes their name, phone number, gender, or whatever.

The human interface is not the world's most challenging either. A few simple
searches, and a way to change the data.

So how does this company manage to screw it up deja vu all over again yet
another time once more still???

There was one big problem with the earlier system. It was hardly ever actually
possible to connect to a server. A person unfamiliar with the DEC way of doing
things might think that this one big problem would be the thing to be fixed
when we produced a new version. 

But NO! We now get three lines of error message instead of one to tell us that
no servers are available! The damn servers are still not available! Damnit,
this is a utility, like electricity! If your electricity is out for 8 hours in
a year, you start muttering about the incompetence of the power company. This
VTX ELF system seems to be out 8 hours a week! If digital was a digital
customer there would be letters to Ken and discussions with lawyers. How many
other CLDs do we have where DEC has managed to irritate almost every computer
user in a multibillion dollar customer? How can anyone at DEC talk to customers
about reliable computing solutions with a clear conscience when we KNOW, from
our own daily experience, that our company does not either know enough or care
enough to make even the most basic network applications reliable? Who is
kidding who about six-sigma quality? Our words and our deeds are so out of sync
its hard to believe the hands and the mouth are attached to the same body!

And suppose the server is actually available. You get into VTX ELF, you type a
command like FIND ROBERT STEWART, hit return, and you get an error message NO
NEXT FIELD. Every other application in the world, you type a command correctly,
hit return, and you get your answer. Not this one! Instead you have to hit
PF1, then ENTER! Not only is it a gratuitously different interface, its
insulting! I know there is no next field, the system knows there is no next
field, the system know exactly what I want and how to get it, but it gives me
some stupid message to prove that VTX ELF is different and I better warp my
brain to like it.

Well OK, suppose you type PF1, then ENTER. What you probably get is no
information about the person's node. This is because that person has not
personnaly gone into ELF and fixed up his entry. First, because that person did
not know that telling the system something it already knew was required.
Second, because that person has not gone to somebodyorother and gotten a magic
password, etc. Damnit, each system on the net already knows who has mail
accounts on it. Poll the systems, get the information, invert it, and you have
a database. Yes, some people would show up more than once, and the system would
have to automatically send mail and ask them what address they prefer. Yes,
some people have the same names and some additional disambiguation would have
to be done. 

But suppose you get past all of this, and actually read what the system knows
about me? Among other things, you would find that ELF thinks my mail stop is
MLO5-2/G10. This is due to some clerk somewhere along the chain mis-entering
the data in one and only one of the two places it appears on the master
personnel form. Actually I am at MLO5-2/G1. I would like to fix this error. I
can't!@!! This damn system goes to all the trouble to give me a secret password
so only I can fix my data, then won't let me do it! I have to find a PSA,
whatever that is, and get that person to fix it for me! Why do I care, you may
wonder? Well, mysterious are the ways of the corporate mail system. Thanks to
that one extra 0, my mail goes to the wrong place. No problem, just walk down
the hall to G10 and get it? No, life is not that simple. There is no G10, so
the company sends all my mail to AUGUSTA MAINE!#@%$^%& No, I am not making this
up.

But I am drifting off the subject. Is it too much to ask a company with a human
interface group, and committees to pass on the minutae of DECwindows style
guides and how many DCL qualifiers can dance on the head of a pin, to produce a
simple human interface that acts like the rest of the company's human
interfaces, and actually lets the user fix wrong information about himself?

Is it to much to ask a company with pretensions to databaseness, to produce a
database that actually has the data in it?

Is it to much to ask the leader in networking to produce a network phone book,
using the information that the network already has, and make it available on
the network?

Is it too much to ask a company with a six-sigma quality group, to produce a
system that is available maybe eight out of ten tries?

Is it too much to ask a company that claims it is going to integrate the whole
enterprise, to just do a decent job of its own phone book?

Isn't there someone in the system integration monster that costs us $90M per
year who would like to prove that the group actually can accomplish something
useful? Here's your chance! Everyone else has failed. Yes, I know that this
problem is more complicated than a software ignoramous like me can code this
evening after work. But if we are going to integrate the enterprise, how about
starting with the network phone book, it will be a great warmup before trying
some real problem.

Sincerely, ^*%#$#@!%$*&&^%#^%$$*^%$^%$@

Robert Stewart

    re .250
    
    What's the secret?
    
    I assume this is the same as the original reply .249 which was also
    set hidden before disappearing into a black hole
    
    j.e.

>                       <<< Note 957.251 by JGO::EVANS >>>
>                         -< Maybe third time lucky!? >-

>    re .250
    
>    What's the secret?
    
>    I assume this is the same as the original reply .249 which was also
>    set hidden before disappearing into a black hole

No secret, just procedures. I deleted .249 because it was set hidden. Setting
it unhidden means that if people have tried to read it already, their seen
maps will have it marked as read and they won't see it unless they specifically
read it by going to that note, rather than using NEXT UNSEEN, etc.

Digital policy requires that I needed the authors permission to post the note.
Even after obtaining this, and saying so in .250, it still wasn't enough for
the moderators, so it was set hidden again. Apparantly, just saying permission
has been obtained isn't enough. Can't moderators just trust people, and 
assume they are telling the truth rather than lying ? The orangebook says
nothing about the moderators seeing the permission (which would also require
permission to forward!!). 

Since the moderator required to see the permission before unhiding the note,
I sent it on to them (without the authors permission, but that's another
rathole). As the moderator in question now has the permission, I have 
unhidden the note. If anyone wishes to read it, it's 957.250

Let's start trusting people, rather than procedures to determine how we work.

	Craig.

>I sent it on to them (without the author's permission)

Permission is only required to post in VAX Notes conferences, not to forward.

I agree.  The new version 2.1 solves almost all the problems I
had with the 2.0 release.  They've done a fine job of listening
to the users.  Too bad it took all that screaming to get their
attention, though.  I'm afraid a lot of folks now have an 
unfavorable impression of ELF/VTX that will take a long time
to erase.

Steve

    I've been in my new position three months.  Since before switching
    jobs, I've been trying to "take all the right steps."  I've notified
    our "CMAS" people about changing my ELF entry.  Before that, I sent a
    DTN book change form to Telecom.  I got an ELF password.  I changed the
    data and was told it would be updated.
    
    I've sent memos to CMAS [I/S & Systems Support] and to Telecom.  I've
    had phone conversations.  I cannot get my old DTN removed from the
    database.  Instead, there are two listed... with my REAL, current DTN
    listed second.  
    
    I give up on that one.  Sometimes getting the simplest thing done takes
    the greatest amount of effort; it gets ridiculous.
    
    P.S.  I also submitted a telecom update form for my new boss to reflect
    her physical move from UFO to UPO and then listed me underneath as the
    "secretary" with my DTN.  The result:  the DTN book lists her new
    address, but her former secretary's number at another facility...  SIGH 

                  <<< JETSAM::ENT:[NOTES$LIBRARY]ELF.NOTE;1 >>>
                         -< Employee Locator Facility >-
================================================================================
Note 200.5               New ELF client for user testing                  5 of 6
I18N::SZETO "Simon Szeto, ISEDA/US at ZKO"           22 lines  13-MAR-1991 17:53
             -< Let's get more participation in this field test! >-
--------------------------------------------------------------------------------
    I would like to point out that field test has started and the field
    test period is only eight weeks.  I think there isn't enough publicity
    around this field test.  Given the brouhaha we had over V2.0, I think
    we should give V2.2 every opportunity for a good workout, or else we
    get to live with the result of non-participation.
    
    I have to confess that I declined being an official field test site,
    for lack of time.  I will use the field test version to the normal
    extent that I use ELF, and report bugs if I find any.  In other words,
    I like to field test V2.2 informally, and I would encourage as many as
    possible of the critics of V2.0 (including those with scathing
    criticism) to field test V2.2, informally or formally.  If this still
    doesn't meet your requirements, here's another chance to speak up.
    
    Presumably it is deliberate that you have to contact Franc Gentili to
    get at the field test kit.  I presume that is to ensure the project
    gets the required number of field test sites.  But this tends to have
    the effect of lessening the exposure of this field test.  You might get
    wider exposure if you just advertised a network kit.
    
    --Simon
    

	We've got a prototype Motif tool called dxelf, which provides
	basic ELF searching capabilities. If you've got a VMS 
	system with the Motif V1.1, or DECwindows V3 FT kit, 
	or an ULTRIX system running V4.0 or later, and a 
	workstation to display on, give it a try. The notesfile
	is KOALA::DXELF.

	Mike