[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::digital

Title:The Digital way of working
Moderator:QUARK::LIONELON
Created:Fri Feb 14 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:5321
Total number of notes:139771

831.0. "AIRING MY VIEW" by DIXIE1::HILLIARD () Tue Jun 06 1989 12:13



	  HOW WE ARE FAILING OUR CUSTOMER'S AND THEREFORE DIGITAL


Did you know that some of our customers feel our products our inferior to 
other manufactures?  Did you know that our customers think our products are 
unreliable?  Did you know that our customers think that some of the 
products we build are lemons?  Where do you think they got these ideas?  
Do you think that some one is maliciously undermining our reputation?  I 
hope to answers those questions and many others in this letter.

We use many other vendor PCS and PDS products in our own computer rooms, 
the DECsite organization uses and recommends other vendors products. This 
leaves the same impression as if we used IBM or Wang for our data processing. 
We are telling customers that our PCS and our PDS are not good enough for 
our own systems. Don't believe it, go look. This is just the tip of the 
iceberg.

Our equipment (any mainframe or hardware device) is designed and 
manufactured to operate with minimum failure in a certain environmental 
window.  From this we base information such as MTTR, MTBF and how many 
individuals are needed to support this device.  We also use this information 
to base the cost of our service contracts.  But guess what, we do not tell 
our customers what that window is, we don't even tell our field service 
representatives.  You say that can't be true, sorry yes it is.  We sell 
customers wonderful systems and the best hardware on the market today and 
don't even tell them what the window is that they were designed to operate 
in. The customer in good faith installs it with no power protection or 
conditioning, temperatures that average 80 degrees and some times with 
chlorine or other harsh chemicals in the room.  Can't happen, surprise, 
it is happening all over the US.  But we give our customers complete site 
preps prepared by qualified individuals, guess again?  They are lucky if 
they even get the receptacle type to plug the device into in time for the 
installation.  So what happens then, the system and it's peripherals run 
like lemons, calls are constantly being logged and we are constantly sending 
someone out there to put good parts in time after time after time. Well this 
probable raises a flag and the environmentalist goes on site and finds the 
problem and saves the day, no, the customer tells all his friends that DEC 
devices are inferior and you know what else, the field service representative 
does not have any background in Environment and he believes that the problem 
is our hardware also, so he stands meekly by and does not even defend his 
equipment. When the environmentalist is called one of two things happens, 
one, the local environmentalist hasn't done a site survey in two years, 
does not remember how and does not have the equipment to do it properly 
because no one was responsible for it and did not keep it up, two, he does 
the site prep identifies all the problems writes a letter to the customer 
gives it to his manager who puts it in the trash, because he is afraid of 
third party maintainnance or he use to work on equipment that was in worse 
shape than this and nothing in either case is done to help this customer.
The next time this customer purchases a computer will it be a DEC computer? 
What do you think he will tell others? How much does this cost us? On
one site in one thirteen week period it cost us $250,000.00 in man hours
alone because they had a bad ground, the ground had 67 amps on it, but
most of the cost is reputation.

But this can't be happening to 8820's, 8830's and 8840's the FIP says they 
will have a siteprep, they are not getting them, they might get a copy of a 
Xsite document that tells them the receptacle type and amount of heat the 
units put out, but not one mention of power, grounding or temperature, 
humidity, chemical contaminations, nothing. Well this is not going to 
happen to our new system, we are going to develop groups of people to 
police this and make sure the customer meets our requirements, remember 
DECsite? We hired EE and mechanical engineers that built aesthetically 
beautiful computer rooms, only one problem it was not designed for a 
computer to function in. We don't need another empire, we have people out 
there that are trained just weighting to be allowed to do there jobs and 
they have something very special, they care and they want to do the job.

From the MVII to the 88's it is not being done.  Why? District managers do 
not feel it is necessary.  It is not in there budget.  They are not 
measured on it.  People are still living in the PDP8 world (did you know 
that this was the first computer to exhibit sensitivity to static, no, well 
neither did our customers or field service people) with less density and 
less speed which made it less sensitive to problems we have now. They have no 
understanding of the ramifications of not doing it until it is generally to 
late and then they still do not remember there lesion and will do it wrong the 
next time. Sells is afraid if the customer knows there is requirements for 
it to operate properly the customer won't purchase it, it will scare him off. 
Sells has asked for field service help in the past and was told we don't do 
that or they asked John or Jane that are not trained in Environment what 
requirements does this device have and in all honesty with no ill intended 
Jane said "Oh, nothing special, at least most of mine don't have any thing 
special and they run most of the time" sells takes this as gospel tells the 
customer in writing and boy do they get in trouble.

Our management has become so number driven they have lost site of there 
goals, customer satisfaction, employee satisfaction and the growth and 
prosperity of this company. They make there numbers at any expense, they 
have told there field service engineers you can not be in the office and you 
can't be on your pager. So what does the engineer do? He logs bogus calls 
and when a real call comes in there is no one to cover it, you don't think 
this is happening I have the numbers to prove it. The number system has been 
so pencils whipped it no longer has any legitimacy, one manager cheated 
to make his numbers this raised the numbers the next time he had to inflate 
them more, etc. Throw out this system and look around you at what is really 
happening. The manager that is not cheating is being punished, if his 
numbers are not as good as the person who is cheating he is told you have to 
many people and the person who is pencil whipping the number system is 
rewarded, does this make sense to you??

What can we do to change this now and in the future?

There is a FIP for Environmental, written by Gerd Wegmann, that could be 
implemented and it would help, it's not the total answer but it is a 
beginning. There are people out there that are trained and have the background, 
ask Dian McClure how many people that "don't exist" attended her meeting. 
Help them get up to speed again. Sells and Field Service have to work as a 
team, if you let a customer know what the window is the system will run 
like a scalded dog and he will want more systems just like it. Measurement 
tools of managers need to be revised to look at this problem and address 
it, if they are not measured on it, it will not be one of there main concerns.  
There are Site Prep packages in existence that have taken a fortune to 
write and publish, the appropriate one for each type of system needs to be 
used.  There is a Installation Management Process Policies and Procedures 
manual that needs to be pulled of the shelf updated and used. We need to 
enforce our real goals not to make numbers but to achieve satisfaction threw 
out our company in an honest fashion. To turn this around would take very 
little effort and money (the cost of not doing it is phenomenal) and it 
would greatly improve customers impression, customer satisfaction would 
increase, manpower would be saved, and parts would not be wasted.  It 
would enable the systems to run as they are designed and boy what a 
difference it will make, we would not be looked at by any one, our 
people or our customers like a manufacture of lemons or a nonprofit 
organization. Employee satisfaction would again be one of our greatest 
resources and assets. We would be number !1!

It has to come from the top and it has to be enforced all the way down the
line to the last individual in the chain. We have every thing it takes to 
make it happen. Even policy's that state it will be done but it isn't (the 
FIP for a 6200 says an Environmental Survey will be done and a siteprep 
given the customer, the 8800's same thing, sells literature given customers 
says they will get this as part of there purchase). This is not an AREA issue 
this is a issue throughout the United States. I have information from many 
other Areas and individuals in those Areas that shows the problem exists all 
over. I first learned of the problem six years ago when I attended a 
Environmental Seminar given by George Yacubovich and Moe Tetrault, where 
they pointed out the hundreds of thousands of dollars wasted each year 
because of lightning, power problems and contamination, things haven't 
changed. The problems are worse as computers are used across the industries 
in areas that they were never traditionally used, yet they have not been 
able to change this because they can only educate they can not make policies 
and see to it that they are implemented. They have tried and met with the 
same frustrations. They could give you example after example, I could give 
you example after example but I do not wish to point a finger at any one 
District or group and perhaps have this cause a loss of focus on the real 
issue, and that is the U.S. has this problem (customers cross all lines of 
Area's in the U.S.)

My motives for writing the letter are that I have been out 
here for a long time and have tried to effectively change what I could but 
the problem is beyond that and outside my ability as an individual to 
change. I do not intend to point a finger at any one group or individual, I 
want our customers and our people to be the happiest and most satisfied in 
the industry. Myself and many others have witnessed the above and can not 
believe it is the intended manner in which this corporations wants to do 
business or the image we want to project.

Digital is number !!!1!!! We have what it takes to make it.




T.RTitleUserPersonal
Name
DateLines
831.1HPSTEK::BURTONFlexmaniaTue Jun 06 1989 13:4532
>                                                                  We sell 
>customers wonderful systems and the best hardware on the market today and 
>don't even tell them what the window is that they were designed to operate 
>in. The customer in good faith installs it with no power protection or 
>conditioning, temperatures that average 80 degrees and some times with 
>chlorine or other harsh chemicals in the room.  Can't happen, surprise, 

I work in High Performance Systems Engineering and design some of the delicate
interconnects for DEC's mainframes.  We make every effort to design and produce
the highest quality, most durable product possible.  Sometimes, however, we are
constrained in what we can do by the current state of technology.  Digital is
constantly asking vendors and researchers to push beyond the state-of-the-art
so that we can continue to design faster, more powerful machines in smaller
packages.  Some of the signal conductors must be as small as a human hair in
order to pack a sufficent number into the alloted space. 

These systems are then subjected to testing that simulates: being dropped from
shoulder high heights by a clumsy delivery person, being left for a week in a
truck in Montana in January, being left for a week in a truck in the Arizona
desert in August, being installed and run for decades, being run on a
production floor that is 80C with chlorine fumes all around, being subjected to
a dust storm, and more.  When we finish, we have a very good idea about what
types of environments the system will survive and for how long on the average.

When people are installing systems, someone should take care that the system is
installed in an environment that is within its specifications.  Those 
specifications are set for a reason.  If we could possibly allow use of the
machine in a harsher environment, we would certainly specify it for such an
enviroment.  The fact is that we have looked at the harsher environments and we
know what will happen.

Jim
831.2VCSESU::COOKVAXcluster Support In-House MusicianTue Jun 06 1989 14:057
    
    re .0
    
    How many times have I said this, READ THE DOCUMENTATION.
        
    /prc
831.3?HPSRAD::KIRKMatt Kirk -- 297-6370Tue Jun 06 1989 14:085
I don't know about other equipment, but my VS2000's manual has, on page 1-3,
environmental requirements and electrical requirements. The requirements
don't specify information about the content of the atmosphere, but they do
discuss temperature, dust, circulation, humidity, static, cleanliness,
storage, amperage, power conditioning, etc.
831.4that's not all .0's storySCDGAT::LINNCitizen of Bullet the Blue SkyTue Jun 06 1989 14:1511
    re: .1, .2, .3
    
    somehow I think you've missed some of the real point of .0
    
    the part about fudging numbers, and believing your own propaganda
    
    re: .0
    
    it's a different world from mine, but, oh, how I understand this
    individual's frustration....
    
831.6DIXIE1::HILLIARDTue Jun 06 1989 14:515
    It is not only my frustration, when you get right down to it is
    is the undermining frustration of this corporation. There is no
    acountability or ownership from the field tech to a long way up.
    We need some strong leadership that is not afraid to make a decision
    and see to it that it is enforced.
831.7CSC32::M_JILSONDoor handle to door handleTue Jun 06 1989 15:2717
From my experience it is the responsibliity of the FSUM to have a proper 
environmental workup done for a new system.  The UM is supposed to recieve 
some communication from Sales when a piece of equiptment has been ordered 
and then she/he must schedule the environmentalist to make a survey.  This 
process admittedly doesn't always work due to communication problems but 
like anything else it works as well as the UM wants it to work.  In my oold 
FS unit we would try our best to have an environmental done before the 
equiptment arrived and were supposed to get one done before installing it.  
We had/have good environmentalist and they did a good job creating a 
package for the customer that contained all the info they need ie KVA, BTU, 
AMPS, recepticle types, cable lengths, etc.  It is up to the UM to decide 
if equiptment should be installed in environmentally suspect locations and 
the responsibility of the FS Engineer to comunicate honestly to the UM 
their evaluation of the location.  I have personally seen it work and I 
have also personally seen it fail.

Jilly
831.8not todayDIXIE1::HILLIARDTue Jun 06 1989 15:412
    Today and over the past seven years it is the exception to do a
    site prep not the ruel. 
831.9happens everywhere...DIXIE1::SILVERSOnsite at Monsanto-Pensacola,FLTue Jun 06 1989 16:486
    Being in SWS, I've had to try and figure out many problems that
    end up being environmentally related, so I share your frustration.
     
    I guess that the root of this frustration is that some of us are
    out here trying to do a quality job, and others are just treading
    water - its very infuriating...
831.10are the metrics correct???PNO::HORNTue Jun 06 1989 17:1412
    .5    It's not so much that there are bad apples in the bunch, it's
    more a matter that we are using the wrong measuring device.  
    It's like measuring a basket of apples by the fluid ounce vs. the
    pound.  ie., We need to develop metrics that correctly measure the
    behavior that we are trying to monitor or improve.  I'm not condoning
    or excusing cheating on metrics, but bad metrics are a form of entrapment.  
    I'd suggest that someone review what needs to be measured (behaviors),
    review the device used to measure that behavior(s) and then take
    the bold steps and change or improve the old metrics.   
    
    Scott
                                                         
831.11Problem partly in the geography....CSC32::S_HALLVAXC SPR == Noop ?Tue Jun 06 1989 17:249
    
    I know the area whereof the author in .0 speaks.
    
    Resources are meager, parts are scarce, numbers are king, and
    often the first response from upper-upper management to a crisis
    is when a customer announces that he's moving the VAXes out, and
    the I*M's in....
    
    Steve H
831.12TRASH THE METRICSKYOA::MIANOWho are the METS?Wed Jun 07 1989 00:4040
RE: .10

>    .5    It's not so much that there are bad apples in the bunch, it's
>    more a matter that we are using the wrong measuring device.  

EXACTLY RIGHT!!!!

>    It's like measuring a basket of apples by the fluid ounce vs. the
>    pound.  ie., We need to develop metrics that correctly measure the
>    behavior that we are trying to monitor or improve.  

COMPLETELY WRONG!!!! Digital urgently needs to trash the "metric" system
altogether.  As long as managers are graded strickly by the numbers
Digital is going to have managers that are good at putting their
interest over Digital's by manipulating the numbers.

First you tell the manager's that they better make their budget or
they're gone. Clever managers then start stomping on customers to get
the quick buck, get promoted, and leave the clean-up to someone else.
To correct that problem you add the customer satisfaction metric.
Everyone in the field knows what a TOTAL JOKE the customer satisfaction
surveys have become.  Managers visit customer sites with the PSS person
in tow.  They tell the customer, politely, "If you don't give us a ten
then I'm going to have to give this Digit a bad review".  Even if
Digital did not do a good job overall, the customer gives good survey
because they liked the Digit and at least HE did good work for them.

When is comes to selling, we tell our customers that they need to look
at an overall solution rather than to pick the computer that does the
best on the Ramp-XYZ benchmark.  We should not be rating people that same
way.  Sure, you can have a budget, but overall performance has to be
graded qualitatively.  The "metric" system is a cop-out and explains why
our corporate vision extends no further than to the end of the next
quarter. 

I could also mention how these metrics are trashing our customer support
(we work at closing out SPRs; not solving problems) but I'll leave that
sore subject to another note.

John
831.13BISTRO::BREICHNERWed Jun 07 1989 08:0114
    I strongly support .12
    Trash the "metrics" !
    Unless someone can show me how to measure:
    Honesty, Doing the right thing, customer satisfaction, quality products
    and customer solutions etc etc...
    After "management by objectives", "management by stats and pie charts",
    how about trying:
    "Management by risk taking" ? Or whatever you call it !
    
    Although .0 elaborates on specific environmental issues, I believe
    having understood the basic message which applies to all of our
    businesses (or at least the ones I know).
    
    Fred
831.14Companies are compelled to measure, but ...VOODOO::VALLETTIJohn Valletti--SOA SIC--AtlantaWed Jun 07 1989 11:5721
	Although it is in the nature of corporations to measure ad nauseum
	it is additionally in the nature of the individual to understand those
	measurements, so that he/she can be successful. This is true no matter
	where your placement is in the organization.

	Measurements drive behavior. So when you have a conflict between what
	is the 'Right' thing to do, versus the 'Measured' behavior, people
	will expend an inordinate amount of energy to try to make both occur.
	Fooling the system takes time, effort ... which convert to $$$.

	When you can alter measurements so that the natural behavior of the
	individual (That is doing things that will make him/her succeed) is 
	also the behavior that makes the customer happy, and the corporation
	succeds, you erase that conflict.

	Will this ever happen ? Probably, ever so slowly over time. You'll know 
	you're getting closer when it's easier to do the right thing for the 
	customer without having to adversely affect yourself or your
	organization. Until then we'll keep trying to make it work.

	-jv
831.15Problem _isn't_ metrics, it's management _BY_ metrics!!!!!SVBEV::VECRUMBAInfinitely deep bag of tricksWed Jun 07 1989 13:3760
re .12

>COMPLETELY WRONG!!!! Digital urgently needs to trash the "metric" system
>altogether.  As long as managers are graded strictly by the numbers
>Digital is going to have managers that are good at putting their
>interest over Digital's by manipulating the numbers.
>
>...
>
>When is comes to selling, we tell our customers that they need to look
>at an overall solution rather than to pick the computer that does the
>best on the Ramp-XYZ benchmark.  We should not be rating people that same
>way.  Sure, you can have a budget, but overall performance has to be
>graded qualitatively.  The "metric" system is a cop-out and explains why
>our corporate vision extends no further than to the end of the next
>quarter. 

I spent a couple of years as a PSS Unit Manager. Let's say my unit budget was
1.3 million in revenue (which the last one was). Let's say I miss by $1. I
fail. I make it by $1. I succeed. Now, remember that margin requirements
require me to make $2 for every $1 in expenses. In every other company, if you
make 25% return on expenses, you're O.K., 50% you're great. Here, 99.99%, you
fail.

The last year I was a PSS Unit Manager, on a 5+ million revenue budget we
missed by about 40K. Guess what? We _FAILED_. (That's less than 1% short.)

The other side, of course, is, say you make TWICE your budget. Does something
special happen? No. You still just succeed. (And, maybe you get to spend a
few days somewhere with other Digits.) 

I won't even talk about the hassles of trying to do the "right" thing while
trying to cope with personnel metrics.

Given the harshness and absoluteness of not making metrics, combined with no
reward (as opposed to "award") structure to recognize outstanding performers,
we're foolish to complain about people's behaviors. And the higher you go, the
the more "quantitative" it gets.

I've been able to grin and bear it, understanding that Digital is perverse and
does not observe any rational business practices as done by any other firm in
the outside world -- certainly none that I have worked for. Perverse behavior,
of course, results from perverse situations. (It's hard to believe that a
company of well over 100,000 employees icould be suffering from some form of
mass hysteria or delusion!)

The problem, however, isn't that metrics "drive" managers or individuals.
The debate isn't over whether metrics are necessary or not (and they are).

The issue is that the people who are managing this company _by metrics alone_*
have effectively, and with surgical precision, excised the "human" part of the
equation. We are now reaping their grim harvest.

/Peters


* It doesn't help, anyway. We announce a wage freeze to contain costs, our
  stock goes down. RJR Reynolds announces they are selling off two subsidiaries
  for 2.5 million to help defray their 3.0 million annual _interest alone_ cost
  from their take-over of Nabisco, and their stock goes up .75 a share.
831.16(slight correction)SVBEV::VECRUMBAInfinitely deep bag of tricksWed Jun 07 1989 13:426
re .-1

>The last year I was a PSS Unit Manager, on a 5+ million revenue budget we
>missed by about 40K. Guess what? We _FAILED_. (That's less than 1% short.)

Oops. That was our district budget. Failing there _really_ hurt.
831.17we can still learn from thisSNOC02::SIMPSONThose whom the Gods would destroy...Thu Jun 08 1989 01:35739
    This was EM'd around quite widely.  Given that it's 25 years old
    the relevance of many of the ideas is quite fascinating:
    
==========================================================================
From:	POOL::ELLIS        "Ron Ellis, VMS Performance"    2-JUN-1989 
14:40:55.90
To:	STAR::FARNHAM,STAR::PROULX,STAR::NIGEL,STAR::GAZO,DOHERTY,HENDERSON
CC:	ELLIS
Subj:	Worsing's speech on reliability and service, a classic

[much forwarding removed]

************************************************************************

IBM used to make a lot of relatively unreliable products...and then they 
changed their ways.  People often ask why.  There are two primary answers.
The first is that early on in the 360 family life IBM only leased systems...
and when a customer was unhappy they simply refused to send in the monthly
payment.  This was surprisingly effective and, as you might guess avoided
a lot of dialogue about "never getting enough data from the field".  Tom
Watson used to get very upset about not getting the rent checks and 
generally did not accept lack of "data" as an excuse.

The other reason is that a few really big customers strongly articulated
their needs in this regard to IBM management.

What follows is a transcript of a speech by R. A. Worsing...then director
of Boeing's computer operation...to IBM Field Service management.
He was invited to speak to them about "service requirements", but
in fact really focussed on product reliability and quality issues more
than on service per se.  Fortunately the speech was transcribed and
has served as a true "classic" within the service community in many
companies.  Note that it was given in 1967, but that many of the points
made are still relevant today within our domain.

The presentation also gives a lot of insight as to why reliability
is an "emotional" arena within any Service community...we all have our
own "Worsings".

Finally, legend has it that for many years UNIVAC required all product
development managers to read this yearly and actually sign an annual
declaration that they had read it.  I cant' verify this, but it wouldn't
surprise me too much.

At any rate, please enjoy and circulate/forward as appropriate.

Regards,
John

******************************************************************************



                        SPEECH TO IBM FIELD ENGINEERING 



                                BRANCH MANAGERS 




                                 July 31, 1967 


                                      By 


                              Dr. R. A. Worsing, 


                                   Director 



                            Systems Administration 

                           and Computing Department 



                              THE BOEING COMPANY 



    TOPICS: 

    1)    What you expect from Field Engineering 

    2)    Importance of System Reliability and Availability 

    3)    Engineering Changes 

    4)    Preventive Maintenance 

    5)    System/360 


  

    My first reaction on receiving an invitation to speak to you this 
    afternoon was one of sheer incredulity.  The few of you who know me 
    well do not, I suspect, regard me as one of your more congenial 
    acquaintances.  "Haven't they had enough of me?" I asked myself. 
    "They must be gluttons for punishment." 


    But being an eternal optimist I came around to the point of view that 
    at last my preaching, cajoling, threatening and pleading had begun to 
    bear fruit.  It will be the first time; but, if I can get my message 
    across, it might not be the last. 


    So I am acting upon the assumption that you are not expecting any 
    bouquets this afternoon.  This is serious business.  The problem, so 
    eloquently portrayed by the movie you have just seen, is bigger than 
    both of us.  Boeing is not a university, it is not an insurance 
    company.  Boeing, because of its size and the urgency of its business, 
    has no trivial problems.  It is a problem amplifier, if you like. 
    Whenever anything goes wrong, it goes wrong badly. 


    Now the area of computing that my people and I are most remote from is 
    precisely the area under discussion today.  I have a lot of 
    programmers working for me and they think they understand software. 
    My hardware people understand logical design and circuitry.  But we 
    have no one conversant with the problem of maintenance.  Consequently, 
    this is the area we least understand and, when you don't understand 
    you become frustrated.  I say this to give you some understanding of 
    the problem of being the manager of the computing department in one of 
    the world's largest, and most rapidly growing, users of computers. 

                                   

    The material which I draw upon for my purposes today has been gleaned 
    over ten years as a victim of outrageous hardware.  I've run the gamut 
    from part user of a 701 to manager of a $25 million-a-year 
    installation.  Today, I have at least one of almost everything that 
    IBM has ever invented.  I suppose God has chosen to punish me here on 
    earth.  For some reason, He just can't wait.  The consequent baptism 
    by fire puts me in an unequalled position to address you today on the 
    subject of Customer Engineering. 


    Carl suggested five subtitles to which I might address myself that he 
    felt would be of significant interest to you.  I am following his 
    advise, and will therefore begin by attempting to answer the question: 
    What do I expect of customer engineers?  Well, there's a one-word 
    answer to that - perfection.  I expect the machines to work twenty-two 
    hours a day.  That's simple enough.  Give me a 100% reliable machine 
    and I will be happy.  Allow one bit to reverse itself in the middle of 
    a two-hour run and I'll be a very unhappy man.  It's simple, home-spun 
    philosophy, attained by bitter experience. 


    If this sounds unreasonable to you, consider the following.  What 
    would you think of an automobile that, despite a thorough daily 
    overhaul, broke down with 15% probability every day?  This is what my 
    grandfather did expect and lived his life accordingly.  However, had 
    his generation found this state of affairs tolerable, our generation 
    would have been in the same boat still today.  They simply adopted the 
    Worsing dictum, you can't abuse all the people all the time.  They 
    finally catch up with you, and you change your ways or perish. 


    Again, how many of you would board a 707 which, despite two hours of 
    preventive maintenance every day, had a probability amounting to 
    certainty that it would require airborne maintenance at least once 
    between midnight and morning on succeeding Sundays?  Or, that you 
    would have to board the airplane twice a day every day for a week 
    or a month to be sure of getting from Seattle to New York once? 

    Let me develop this idea by putting into Boeing's mouth the kinds of 
    repartee that I get subjected to.  In this way, you might understand 
    more readily and be more sympathetic to my cause.  Now I regard a main 
    frame, the card readers, printers, tapes, discs, drums, telephone 
    wires, modems and remote I/O units as a single system.  If any device 
    is down the entire system is down.  There's no such thing yet as 
    "partly down".  Computers are binary, aren't they?  This means they 
    are either working perfectly or they aren't working at all.  What 
    comfort is it to the programmer to be told that his program almost 
    ran,and had it not been for a loose connection on chassis five, he 
    wouldn't have been here at three in the morning?  Now let's change 
    roles. 


          "Take courage, Widow Robinson, it was only the engines that 
          conked out.  The body and wings were in perfect shape.  Anyway, 
          we don't make engines so it wasn't really our fault." 


          "Members of the Inquiry Board, the plane behaved perfectly all 
          week.  Just a loose connection with the main fuel tank in the 
          middle of the Indian Ocean for three minutes.  It had passed all 
          the pre-flight diagnostics." 


          "Of course, we don't guarantee a thing.  If you want to be 
          absolutely certain of getting from New York to London, you'd 
          better send three together and transfer the passengers in 
          mid-flight as necessary." 


          "Ladies and Gentlemen, we request your patience at this time. 
          It should only have taken twenty minutes to change wings. 
          However, owing to an absolutely unthinkable and unforeseen 
          problem, we can't get the new ones on properly.  If you would 
          kindly disembark, you will be taken to the passenger lounge 
          where complimentary instant coffee will be served." 

          "Ladies and Gentlemen, owing to an utterly mystifying series of 
          events, our stock of spare hub caps has been found to be 
          exhausted.  However, a spare is being flown in from Karachi on 
          Thursday.  We do want you to know that this sort of thing 
          doesn't usually happen because we maintain stock levels that are 
          based on years of experience.  We know you find this comforting 
          and hope you will find your enforced sojourn in Thule to be an 
          unexpected pleasure." 


    Well, why is this so ridiculous?  Why do you expect so much more 
    reliability with an airplane than with a computer?  They aren't any 
    more expensive.  They are much more difficult to pilot.  They are not 
    kept in a special environment. 

    I say that the only reason is that airplane customers regard degrees 
    of reliability as unthinkable, while computer customers do not. 

    And the reason why they don't is that, they've been knocked on the 
    head so often, they are incapable of clear thought on the matter. 
    Well, my head is made of concrete, and my vision is quite clear.  It 
    is just as intolerable to allow the collection of gadgets we call a 
    computer to admit of unreliability, as it is to allow the collection 
    of devices we call an airplane to admit of unreliability. 


    So, to repeat myself, I expect from you simply - perfection! 


    Anticipating the howls of protest, I want to bring out the following 
    point.  It is a point which is not common currency, and one which you 
    may not recognize, or agree with, because you are too close to the 
    situation.  It is this:  The relationship between the field and the 
    plant is not conductive to satisfactory performance.  To put it more 
    plainly, you are not being supported properly by your people back at 
    the plant.  And while I regard you as partly responsible for this, I 
    really think that the people back at the ranch - Poughkeepsie, 
    Boulder, etc. - are mainly to blame. 


    As far as I am concerned, you are the people responsible for 
    reliability.  I can't afford to take any other point of view.  When 
    the machine is down, you are to blame.  But, privately, I recognize 
    the fact that you can't make a silk purse out of a sow's ear, and if 
    Headquarters is in the pork business instead of the silkworm business, 
    then you don't stand much of a chance.  Privately, I recognize the 
    fact that reliability is a function of good design, manufacturing 
    excellence, quality control, etc.  However, you will never get me to 
    admit that on Boeing territory, precisely because I cannot allow 
    myself to get into the position of holding more than one person 
    responsible for reliability; or for anything else for that matter. 

    I have to force the system to work.  I refuse to crutch it by going 
    directly to your Engineering or Manufacturing people.  Anyway, your 
    company won't let me.  So I go to you and I demand perfection.  It is, 
    in turn, your responsibility to see to it that the machinery you 
    service is designed and fabricated properly.  It is this step that I 
    find lacking.  My experience has indicated to me that there is a very 
    poor relationship between home base and the colonies. 

    I don't say that Headquarters doesn't get to hear from you.  You wrap 
    every defective part in a pre-addressed box and ship it into the wide 
    blue yonder.  From a statistical point of view, Headquarters knows 
    what's going on quite accurately.  I don't doubt that.  What I do 
    doubt are two things, namely that you receive anything back from them 
    in the way of insight, visibility, statistics, etc., and that they 
    receive anything from you in the way of semantics.  By the latter, I 
    mean for example, they receive a defective five-cent console 
    typewriter connector - just like I'm wearing on my tie clip - and they 
    say, "This makes only 943 defective five-cent connectors this week. 
    We're down 7% on last week.  Keep up the good work, lads."  What they 
    don't receive is the information that has caused: 


          a.   the entire system to be unavailable for four hours at the 
               cost of $2,000; 

          b.   the rerunning of a two-hour job for $1,000; 

          c.   the delay of all scheduled work to the factory that day, 
               causing delays in the factory at the cost of $200,000. 


    How often have you ever told that to the factory?  How much vinegar do 
    you pour into the pipe-line?  Statistics, unaccompanied by semantics, 
    are a cold and lifeless thing, and your job is to breathe the breath 
    of life into your reports to awaken the Rip Van Winkles back home.  In 
    the meantime, my job is to render every assistance I can by pointing 
    out your shortcomings. 

    The second subtitle concerns itself with the importance of system 
    reliability and availability.  Obviously, the foregoing was 
    inextricably mixed with this subject because, really, what I expect 
    from Field Engineering is precisely that product-reliability and 
    availability.  However, there is a bit more to be told. 


    Firstly, I want to restate the fact that Boeing is committed to 
    computers.  There's no way back.  As computers became available, and 
    as we learned how to use them, we slowly hitched ourselves to their 
    tremendous power.  But, we didn't just replace the machines.  Rather, 
    we started doing our business in new ways.  We began to get used to 
    expecting new kinds of service.  New relationships got under way.  And 
    all this novelty was characterized by orders of magnitude of 
    difference in volumes of information, timeliness of information, and 
    reliability of information - both on the scientific and the commercial 
    side.  We can now do stress analysis using so many points that squads 
    of people would take lifetimes, by hand, to carry it out; and we can 
    do this overnight. 


    We process all the paperwork which the factory needs to bring the 
    parts to the right places at the right times, overnight.  The orders 
    for the day are out to the troops, before the troops come aboard every 
    single day.  And the orders are accurate.  The size of our final 
    assembly operations absolutely precludes the possibility of doing this 
    by hand.  If we were making one product with little change, it 
    wouldn't be such a problem, and everything would be part of a routine 
    cycle, but just about every airplane is different to its neighbor.  We 
    don't mass produce, we custom-build. 


    Now the great coincidence - a coincidence that is largely not 
    understood, even by many of our people - is that computers became 
    available just when they were wanted.  You see, you can't do the 
    stress analysis of a 707, 747, or SST, without a computer; and you 
    can't have a customer-oriented assembly line without one either. 


    Without the computer, airplane development would have come to a halt. 
    Neither can you have a Space program, or build thermo-nuclear devices. 
    This, I think, is a fundamental fact of modern life largely overlooked 
    by most people, including, I regret to say, the computer manufacturers 
    themselves.  That's why the sales talk still stresses 
    people-replacement and economy of the old thing, instead of capability 
    for the new thing. 


    So you see, there's no going back.  We are committed.  In order to 
    build airplanes you need people, buildings, tools, materials - and 
    computers.  And to weld these attributes into a viable unit, you must 
    plan and manage that plan, and this entails commitment.  And to commit 
    a computer is to state, categorically, that it shall function. 

    I hope you find this a simple message. 


    Secondly, I want to take up cudgels with your criteria.  I want to 
    discuss, for a moment, how you should be measured and how you should 
    be reported.  What is availability? 

    Each week at our IBM-Boeing Review Meeting, and each month at our 
    meeting with Buck Rodgers, we discuss the incidents of downtime, how 
    long devices were down, what went wrong, how you located the trouble, 
    what you did to fix it, and what you are going to do to prevent it 
    recurring.  All of this is necessary, useful activity, but it stops a 
    shade short of reality.  The quality of your work is displayed in 
    terms of graphs showing "availability" - and I say this in quotes - 
    and as this quantity increases by percent, we all sit back deluded by 
    the appearance of improvement, basking in the glory of our diagnostic 
    prowess.  Sometimes the graph of tape unit performance shows 99.9% 
    availability.  It sounds like a Russion election, and in a sense it 
    is, as I shall explain. 

    How come that, in a week of superb equipment "availability", it is 
    possible for our applications people to curse your very bones for 
    causing them daily delays in schedule?  Are they getting at the wrong 
    guy?  It doesn't seem justice does it?  But, the way we report things 
    today, it is only too possible.  You see, in a week of "high 
    availability", it isn't the total length of down time that causes the 
    trouble, it is the frequency.  If a tape unit goes down, I don't 
    really care how long it is down, provided the others continue to 
    function.  If it's down, get it out of the way and don't bring it back 
    until you can guarantee it.  So it can well happen that "availability" 
    is down but Worsing is "happy" - I put that in quotes, too, you 
    understand. 

    What causes the trouble is the eye-blink downtime in the middle of the 
    two-hour run.  The eye-blinks don't integrate into anything 
    displayable in charts each week, but the consequent reruns can add up 
    to disaster.  So the true availability is total time, minus rerun and 
    delay time.  That's the way you should really be judged. 

    There are some complications, of course.  You can complain that if you 
    luck out, the eye-blinks will occur in two-minute runs so that total 
    rerun and delay time is not the measure.  Perhaps frequency is a 
    better measure. 


    Then again, it is difficult to obtain some of the numbers.  Tape 
    parity errors, for instance, to state the richest source of our 
    displeasure.  I will be dealing with this problem in detail at the end 
    of my talk. 

    Whether you can see all the answers or not, at least I hope that you 
    can see the need for reliability and the need, accurately, to portray 
    availability. 


    Now I come to the thorny subject of engineering changes.  Engineering 
    changes are regarded as necessary for four reasons, namely: 
    capability, error-correction, reliability, and serviceability.  I'll 
    take each in turn. 

    As far as capability is concerned, and I take this basically to mean 
    speed, I am not interested.  Don't waste computer time speeding up the 
    circuitry.  I ordered the machines according to a schedule of speed 
    specifications and I'm content with that.  The only exception to that 
    general rule is where you have a design mistake, and the machine has 
    timing troubles.  Of course, you have to fix that, but that comes 
    under error-correction. 


    If you want to speed things up significantly, put all the E.C.'s on in 
    the factory, give the machine a new name and replace the one in house. 
    Then I'm never down. 

    Regarding error-corrections, these, of course, have to be made; and if 
    they are few in number, please install them promptly.  If they are not 
    few in number, then I am the possessor of a citrus fruit and I want 
    neither peel nor pip of it.  Get it out after you've got its 
    replacement up and running, and don't have the effrontery to ask for 
    any rent. 

    I think by far the greatest E.C. activity is in the realm of 
    reliability.  I think you can always find ways of making the machine 
    more rugged.   When you build a new model, you like to use the most 
    advanced technology available, consistent with cost and risk.  And, by 
    definition, this means technology untried in the field.  And, in turn, 
    this implies a more than marginal propensity for trouble, particularly 
    in the early stages. 

    It goes without saying, then, that reliability improvements are going 
    to be rife in the first year to any model.  We have seen this quite 
    dramatically in the case of the model 75.  When it first hit the floor 
    it was, frankly, a totally useless machine.  Its central circuitry was 
    just not up to the trask required of it.  But, since the repopulation 
    in March, it has become perhaps the best of all the 360's despite its 
    speed.  Again, in the case of the 91, you threw some $24 million worth 
    of parts into the Hudson - that's a pretty expensive form of pollution 
    by the way - for the same reason. 


    But when should the stream of reliability E.C.'s begin to dry up?  I 
    ask this, not as a rhetorical question, but for understanding.  I 
    really don't know, and I think a frank statement of IBM's goals and 
    expectations with respect to the 360 is about due.  The big machines 
    have been out for over a year now, and I think IBM should know enough 
    to be able to tell us. 


    When you design a machine for a new technology, you don't know what 
    the problems are going to be, so you don't know what troubles you are 
    going to have fixing the problems; hence, the fourth type of 
    engineering, change for serviceability.  If the only way to change the 
    hammers is to hang upside down on a rope from a pully in the ceiling, 
    then you obviously have to install an E.C., ie, you have to strengthen 
    the ceiling. 

    Serviceability means more diagnostics in the allotted preventive 
    maintenance period, and less time needed to locate and repair faults. 
    So, let's have serviceability engineering changes. 


    Having covered the reasons for E.C.'s, I now want to mention the vexed 
    problems of scheduling the time needed to install them.  What are the 
    premises?  Certainly, if the need for improvement is recognized, the 
    improvement should be incorporated.  But, if the machine has been 
    working reasonably well since the need was recognized, one more day 
    won't do it any harm. 

    Parenthetically, if it hasn't been working well, then the E.C. is not 
    regarded as such, but as part of a program to get the machine up and 
    running.  Furthermore, it takes time to install changes, and I don't 
    want to use more time in installation than I save by installation. 
    And I want preventive maintenance time taken entirely for preventive 
    maintenance.  I can only allow E.C. installation provided the standard 
    crew performs its  normal procedures without any interference.  In 
    addition, E.C's, by their very nature, change the machine to something 
    different, and the consequences of this cannot be predicted with 
    complete accuracy; therefore, I cannot be certain that the machine 
    will recover from the trauma. 

    This is a pretty formidable set of conditions and restraints, and I 
    have no formula for success.  What I think I would prefer to do is 
    give me: 

          a.   A plan of philosophy and expectations for each machine, 
               updated as experience is acquired. 

          b.   A statement of the risks involved in allowing E.C.'s to 
               accumulate. 

          c.   A general plan for acquiring the time to install the main 
               bulk of the E.C.'s, and 

          d.   A coordinate schedule for the installation of E.C.'s 
               requiring lengthy installation, or having high probability 
               of causing subsequent morning sickness. 

    That concludes all I have to say concerning engineering changes and 
    leads me to the next topic, that of preventive maintenance. 

    I understand the need for preventive maintenance, and am happy to 
    schedule a period each day for its execution.  This is supposed to 
    purchase my ticket to perfection, reliability, availability, and all 
    stations to happiness.  What you do with your preventive maintenance 
    time is not regarded as any affair of mine.  I put on my walking shoes 
    and I sneak past squads of engineers armed to the teeth with 
    ringbinders and scopes, but I never intrude.  And if all were well, 
    I'd stay clear out of the way.  But it isn't.  How many times does a 
    device that has been running perfectly all week, crop out within the 
    first hour of preventive maintenance on Friday?  And not just trivial 
    devices, either.  How often has the two-hour preventive maintenance 
    period been followed by a four-hour post-p.m. recovery session? 

    Now, I'm the first to admit the existence of Heisenburg's Principle: 
    You can't measure the temperature of the bath water without changing 
    it, and I suppose you can't test equipment without imposing some 
    strain on it.  But sometimes I think you're too brutal. You seem to 
    split the patient's head to find out if he has any brains. 

    But it seems computers are designed for smooth running, not for 
    diagnosing.  Perhaps we need a diagnostic-oriented computer.  We 
    already have primitive diagnostic circuitry in some of the 360's. 
    Perhaps this is the first step towards the fail-softness you keep 
    telling me is just around the corner. 

    But today I get the very definite feeling: a) that the machine hates 
    being tampered with and, b) that your techniques aren't sufficiently 
    scientific or comprehensive. 

    Why should we have better diagnostic programs than you?  But we have. 
    By your own admission, there's no better diagnostic than a production 
    program, but you can have a copy of that program any time you like. 
    You have access to everything we've got, plus your own staff, so 
    really there's no excuse for handing over the machines to us without a 
    guarantee of current perfection, is there? 

    What is haunting both of us, I think - and here I find myself banded 
    with you against our common enemy, the Poughkeepsie Business Machine 
    Company - is that with today's machines, it is very difficult to tell 
    sometimes whether the machine is up or not.  Tapes may be spinning, 
    messages frantically typing themselves on the first generation console 
    typewriter - they're so cryptic anyway as to be indistinguishable from 
    random sequences, lights flashing, yet nothing but chaos being 
    created.  Another way of expressing this, perhaps, is that production 
    work looks pretty chaotic.  We're OK when we get a hard stop.  All 
    systems are stop, we fill out a card, hand it to the C.E.'s, and go 
    off for a smoke.  But IBM has invented a thing that I hereby name the 
    "soft stop".  It's dead, but gives all impressions of being alive - 
    like some people I know.  It takes very clever operators, assisted by 
    Knowledgeable software people often, to issue the death certificate. 


    What's it going to be like with multiprogramming?  Have you considered 
    this problem in the multiprogramming environment? 

    My mind boggles at the contemplation of the interaction of coexistent 
    programs on a soft machine.  are you equal to the task?  Or are you 
    going to be a reincarnation of the Grand Old Duke of York? 

    Well, to summarize my thoughts on preventive maintenance, I will 
    merely repeat that I'm in favor of it, but I feel that there is still 
    plenty of scope and requirements for improvements in your use of the 
    time. 

    My last subtitle is what I think of the 360.  This, of course, is 
    really a speech in its own right.  I don't have time to give more than 
    a brief summary here today. 

    Firstly, I think it is appropriate to state my reasons for choosing 
    the 360 in the first place.  We chose it for its peripherals. 
    Peripherals are the key to data processing, not main frames.  I should 
    point out here that I do not regard the 360 as a computer, and for 
    that reason, we do not use it as such.  We do our computing, our 
    arithmetic that is, on the machine of another vendor.  The 360 has a 
    short word-length.  It is a hexadecimal machine which makes it 
    effectively even shorter, and it does not have floating-point round. 
    You can say that the model 91 commits round-off error faster than 
    anything on the face of the earth, while the first lecture in the 
    introductory course of any Numerical Analysis Program states that this 
    is the fundamental problem of computation, and the source of all evil. 


    It is one of the curious mysteries of IBM, the omission of rounding in 
    the 360.  It would be less of a mystery if IBM were to come right out 
    and say that the 360 is intended for data processing and not for 
    computing. 


    However, back to the mainframe theme.  Data processing is essentially 
    the business of obtaining correct bits, storing them on something for 
    periods of time, and being able to read them from storage whenever I 
    like.  Questions of language and speed are secondary.  The question of 
    reliability, as you know by now, is the primary one. 

    As you also probably know, the most unreliable device in data 
    processing is the conventional half-inch magnetic tape.  This has 
    caused The Boeing Company's Commercial Airplane Division more trouble 
    by an order of magnitude than anything else.  It suffers from problems 
    of environment (dust particles), electronics (it is highly analog 
    rather than digital), mechanical wear (oxide on the tape), mechanical 
    obstinancy (reels spinning in opposite directions).  Half-inch 
    magnetic tape, even today, is a Rube Goldberg device, very little 
    removed from the string-and-pulley device of fifteen years ago.  Yet 
    we continue to commit the invaluable information of the company to its 
    rugged surface.  You might just as well put to sea in a sieve! 

    So that was the first problem to solve.  We had to get off the 
    conventional half-inch magnetic tape onto some sort of twentieth 
    century device.  And if this were all we achieved by going 360, it 
    would still be a significant improvement to the business of designing 
    and building airplanes. 

    We looked at all the gadgets being offered by the manufacturers, and 
    we formed the opinion that hypertape and the 2314 were the things we 
    needed to replace the 729.  So, quite categorically, I say that we 
    embarked upon the nightmare of conversion to get our hands and our 
    data on hypertape and 2314's. 


    But, when you stand back and survey the scene in perspective, why did 
    we have to change so much else just to do that?  The only admissable 
    answer to that is a long term one; you want these devices usable with 
    a lot of other devices, and the 7080's and 7094's couldn't begin to 
    support them.  So you had to have a more sophisticated device to 
    operate these peripherals, and we call that device the 360.  However, 
    as a short-term proposition going from 36 to 32 bits, binary to 
    hexadecimal, commercial translator to COBOL, control cards to Job 
    Control Language, is proving a traumataic experience, and in the year 
    that we've had the heavy-frame 360's, we have made no discernable 
    progress in unloading the 7080's.  And some people have even had the 
    misfortune of following IBM's advice to convert to PL/1.  I think 
    that's about the only mistake we didn't make. 

    It hasn't been easy, but at least we are now using hypertape and 
    2314's on a daily basis and, so far, these devices have lived up to 
    our expectations.  We are cycling the hyper cartridges through a 
    testing and stripping operation on a monthly basis.  We don't adtually 
    strip them like we do 729 tape, we merely punch a new hole further 
    along the tape. 

    One somewhat puzzling fact that might be of interest is that, as far 
    as I know, there are only two installations that use hypertape, namely 
    the Internal Revenue and the Commercial Airplane Division of The 
    Boeing Company.  Aren't other people having the same trouble with tape 
    as we are?  Isn't their data as valuable?  Or is it that they don't 
    have such critical deadlines as we do and can afford to rerun. 


    To the Internal Revenue, on the other hand, information is everything. 
    The value of the data far exceeds the additional cost of hyper, and 
    they have paid whatever it takes to get the best reliability 
    available, and that's what we have done.  The fact that no one else 
    has does not worry me unduly.  We've always led the world at Renton, I 
    regret to say. 

    One aspect does worry me though, and that is IBM's understanding of 
    what it is doing.  Hypertape is the principle reason for going 360 at 
    the world's largest computing center, yet IBM doesn't even support 
    hyper.  We have dragged hypertape, kicking and screaming, on to the 
    360 and the only help we get from IBM is local support.  None of the 
    versions of OS/360 that leave Poughkeepsie have the vaguest awareness 
    of hypertape, so we have to glue it on ourselves at Sys. Gen. time. 
    This is a particular example of the general malaise in IBM and, I 
    think, in the other manufacturers - they don't know what computers are 
    for.  They have little understanding of applications, even less of 
    software, less still of the concept of total systems.  I'm still 
    uneasily suspicious that, to the manufacturers, a better computer is a 
    faster C.P.U.  This is certainly true so far with respect to the 91, 
    and that is why we have deferred and modified our order for two.  No 
    consideration at all has yet been given to the problem of how to get 
    data into, and out of, the computer.  The sole design criterion of the 
    91 was a fast arithmetic unit.  This worries me.  It worries me 
    greatly.  It undermines my confidence.  "What are they going to do 
    next?" I ask.  You must remember, we have very little control of our 
    computing environment. 

    All we can do is take what the manufacturers offer.  We do not have 
    the surplus to design computers or write software.  We have no 
    influence over those who do.  We pay out the equivalent of four 707's 
    each year in rental, but that does not buy us one cent's worth of 
    influence or control.  My company is absolutely, irrevokably, 
    committed to computers, but it has very limited voice in what a 
    computer shall look like, and no assurance at all that its needs of 
    the future shall be satisfied.  For example, right now, we just can't 
    get 2314's, and there are no packs available for the 2311. 

    To return to the positive aspects of the 360, our second reason for 
    choosing it was that it would enable us better to map over the 
    operations in the factory to the operations in the computer.  The 
    factory is a continuous thing, but all we've been able to do hitherto 
    is to model it in a very discontinuous way.  We would punch up a lot 
    of cards and run them on  the machine once a day.  It's like taking 
    one breath a day and hoping to stay alive.  But that's all you could 
    do on a large scale on the old machines.  Now we have the capability 
    of designing systems that will produce the results that the factory 
    needs on a time scale that suits the factory, rather than one that 
    suits the computing department.  So we have begun installing model 
    20's and 30's as remote card readers and printers, alongside the 
    1030's and 1050's already hooked to the 1460's.  We have just started 
    to use the 2321 Data Cell Drive, but it's too early to say anything 
    about that yet. 


    In time, we will know how to use this rich array of devices.  It will 
    no longer be necessary to prescribe leeches for every ailment, instead 
    we'll let the punishment fit the crime. 

    We'll know enough about the economics.  We hope we'll be able to find 
    out how many to order of a particular device to ensure the 
    availability of one.  We hope that our programmers will have mastered 
    the subtleties of Job Control Language.  But, when the day of 
    revelation dawns, please don't announce System/7.2832, the 9-bit 
    bytes, PL/2, Supertape, Dinky Disk, Data Jail, excommunication devices 
    and voice input - however trivial the conversion.  We have enough to 
    keep us busy for the next decate, and all we need from the 
    manufacturer is unprecedented availability.  You, the purveyors of 
    availability, are now in the forefront of the struggle to make this 
    vast complex of possibilities a viable, economic reality.  I am sure 
    you are equal to the challenge, and I am confident of your continuing 
    ability to ensure its eventual success. 

    I thank you deeply for the opportunity of conveying these thoughts, 
    impressions, criticisms, questions and aspirations which I hope have 
    been taken in the spirit given, and would be happy to engage in 
    vigorous debate or discussion if there are any questions from your 
    side. 
831.18site prep on VAX 6000?LANDO::NOYESCalypso EngineeringFri Jun 09 1989 15:0314
    re .0
    
    I was the responsible person for developing site prep data for VAX 6000
    systems (Calypso).  To make the argument more specific, do you have any
    particular issues with Calypso site prep? We tried to clearly outline
    it in the Systems and Options Catalog and in the Installation Manuals.
    Did we get it right? If there are additional specs or parameters that we
    should be supplying please let me know and we can see if its worthwhile
    to add it in. I haven't heard any complaints (yet) about what has been done.
    
    The SOC is the right place for it. It goes to everyone, customers and
    the field.
    
    Steve
831.19DIXIE1::HILLIARDFri Jun 09 1989 15:102
    The FIP indicates one is required, but I know it is not happening.
    Some have had problems as a direct result of this. 
831.20bit me in the a__SALSA::MOELLERDistribution list mercifully deleted.Mon Jun 12 1989 14:366
    re .18, site prep for 6000 series.. it would have been nice if there
    had been a big RED FLAG regarding the requisite VMS version (5.1)
    for the 6310.  As it was, all I found was one sentence buried in
    a Sales Update article.
    
    karl
831.21LESLIE::LESLIEAndy =^= LeslieTue Jun 13 1989 05:476
    
    I suggest you mail the author of the Sales Update article pointing out
    this deficiency and ask that it not be repeated. Moans here may not fix
    the problem.
    
    - Andy
831.22plus ca change, plus ca memechoseCHEFS::CONWAYWed Jun 14 1989 05:4422
    In the six days since .17 was posted the only reponse seems to be
    what a wonderful/lousy job we did on the 6000 series environmental
    spec - which is exactly what the note implies is the problem.
    
    Do we believe .17 doesn't apply anymore ? or apply to us ?  Whether
    or not the evironmental spec was correct, adequate, or published
    correctly is only a symptom of the problem.
    
    I suggest that what Boeing were finding in the 60s applies to many
    more customers today - when their machine stops so do they.
                                                            
    We have to get our total act together so well that every machine
    works first time, every time, in every situation that we sold it
    to them for.  And when it doesn't then there is only one metric
    that matters - getting it working again as soon as possible
    commensurate with the customers needs for that system. That
    is doing the right thing.
                                                     
    As the paper implied we don't expect the flight engineer to be
    adjusting the engines at 35000' ( even on a 737-400 ! ).
                                                     
    
831.23I keep on tryingODIXIE::HILLIARDMon Jun 26 1989 10:177
    I entered this note out of frustration and the hope that some one
    would read it that had the authority to do something about it. I was
    wrong. There is no accountability and there for no one is responsible
    for the problems. So it stands to reasion no one can correct it.
    I will continue to try in every way I can, every chance I have.
         
831.24TRCO01::FINNEYKeep cool, but do not freeze ...Mon Jun 26 1989 21:4125
    >>I entered this note out of frustration and the hope that someone
    >>would read it that had the authority to do something about it.
      
    If that is why you entered the note, then you did it for entirely
    the wrong reasons.
    
    >>There is no accountability and there for no one is responsible
    >>for the problems.
      
    If you make this statement based upon the response received in notes,
    then you have reaced an incorrect conclusion.
    
    >>So it stands to reasion no one can correct it.
    
    Wrong, based upon above.
    
    >>I will continue to try in every way I can, every chance I have.
    
    Good, but forget trying to rectify any policy, or corporate strategy,
    through notes. That is not the mandate, nor the function of notes.
    
    Direct your energies at the appropriate places. If you can't find
    the target, start at your manager, and work up.
    
    Scooter
831.25ODIXIE::HILLIARDWed Jun 28 1989 10:0918
    Frustration, in as much as I was shouting and that I felt the people
    in the note would understand. 
    
    Accountability is a real issue and a large problem, this is deducted
    from being out here and interacting with four other Areas.
    
    I did not expect any one reading this note to change any thing major
    but I do expect if it pointed out things they were not aware of
    that the individuals would investigate and change what they could.
    
    I am changing the things that are in my power to change and I always
    address my local management and give them my information and ideas.
    They are usually very receptive and they are working to improve
    a number of things at this time. They read the letter back in Jan.
    before I ever entered it here. But as I said this is not and individual
    Area issue.
    
    
831.26Do you feel better now?DR::BLINNThe best mechanics are self-taughtTue Aug 22 1989 19:4725
        RE: .0, .25 -- Posting a note in this conference may be a good
        way to vent your frustration or get your name in lights, but
        it is NOT a way to fix a problem.  The note you posted clearly
        seems to have been written to address some specific problems,
        and must surely have been intended for someone specific, but
        quite frankly, it's impossible to tell (without more context
        than you've provided) who is impacted, who is supposed to fix
        the problem, and who you've told about it.  The advice in .24
        is exactly correct.  If you are aware of a problem and you
        want it fixed, don't write a note in a conference like this.
        Take personal responsibility for the problem, and fix it. 
        
        On the other hand, if all you want to do is vent your feelings,
        writing a note here is probably as gratifying as other things
        you could do.  But it probably won't fix the problem.
        
        As for reply .17, it would have been nice if the person who
        transcribed the copyrighted material (which, if I remember
        correctly, was published in the Annals of the History of Computing
        a while back) had correctly attributed it.  It would be even
        nicer if people wouldn't post MASSIVE notes like that.  If
        there's relevant material, summarize it and point the world
        to a copy.  Please.
        
        Tom