Conference 7.286::digital

	But the article doesn't say how the demonstration was received!!!!

    Well, our stock went up 2.  But given Wall St. the two may have no
    relationship to each other anyhow!
    
    /Charlotte

    What is the stock at anyway.
    
    
    		Kirk.

   
   ...they all rushed out to sell their DEC stock and buy MicroSoft. :-) :-)

    It seems like these days that when the DOW goes down, DEC Stock goes
    down. When the DOW goes up, DEC stock goes down anyway. When DEC doesn't
    announce layoffs, DEC stock goes down. When it does it goes down anyway
    because the numbers are not enough. :-(

    Nari 

    

    

    DEC 34 1/8, change -0 7/8; DJIA 3253.02, change +1.62 at 12:22

    I was there in the audience.  Bill Gates was very enthusiastic about
    Alpha.  "Windows on Wall Street" is a Microsoft marketing event for
    users of Microsoft products, not a presentation to industry analysts
    although they may have been in the audience.
    
    Gates had nothing negative to say about Digital, and clearly positioned
    Alpha as superior to Pentium and other RISC processors.

    
    That article left me really cold. Just another thinly veiled
    advertisement. It said nothing about how either Windows NT or Alpha
    ACTUALLY performed. It just spit out the same old stuff about how they
    are so great etc. Don't get me wrong here, I hope Alpha becomes the
    next thing that everybody MUST HAVE. But I'm kind of tired of hearing
    how great it is from DEC marketing people, I'd like to start hearing it
    from people outside of DEC who don't have any interest at all in how
    well it sells.
    
    Steve B.

    I beg to differ!  It says a lot when your hot new product performs well
    under the hand of a top-level manager ...  extra points if (s)he doesn't
    get his/her hand zapped in the proess ...
    
    Steve

    re .8
    
    The article mentioned NOTHING about how well (or not well) the demo
    went. So maybe the top-level manager DID get his/her hand zapped.
    
    Steve

re Note 2186.8 by ECADSR::SHERMAN:

>     under the hand of a top-level manager 

        Besides, Gates is richer than Perot and just might run for
        president some day!

        :-)

    The reason it says nothing about how the demo went is that it was
    written the day before for release "as it happens".
    
    ...tom

    
    re .11
    
    Oh, sorry. I thought it was supposed to be a Wall St Journal article. I
    just looked back at the base note and found no mention of the WSJ. Oh
    well, if it's a press release, I guess it does what it is supposed to
    do.
    
    Nevermind,
    
    Steve B.
    
    P.S.
    
    I'm really getting sick of that tired paragraph about Digital being
    "The largest supplier of Networked etc. etc." don't they have any
    imagination? They could at least change the wording once in a while. It
    looks like they just paste the same thing into everything they send
    out. I just had to nit pick a little :-).
     

    re .12
    
      That blurb is a so-called 'motherhood' statement and is used exactly
    as you suspected.  To my knowledge it was last revised about 2 years
    ago.
    
      Imagination is a quantity that is not always highly valued here, or
    most large institutions for that matter.  Uniformity and lock-step
    adherence to the corporate line seems to be the desired trait.  I hope
    the new guard will light a fire under our marketing/advertising teams
    soon.  If not, even if ALPHA turns out to be the 'must have' product of 
    the 90's no one outside of Digital will know it.
    
    Marv
    

    An earlier note voiced the wish for some 'good words' from someone
    besides DEC-insiders...
    
    Take a look at the current week Computer World. There is an article on
    page 107 (i.e. buried from sight) which quotes the users at the Mayo
    Foundation who ported 400,000 lines of VAX code to an ALPHA in a mere 3
    weeks. They even said that they could've done it in 1 week with some of
    our improved tools now available.
    
    THIS IS GREAT NEWS!  How many System 36 managers who barely survived
    porting to AS/400 would love to be able to tell that story? For that
    matter who has tried porting from PDP-11 to a VAX with that level of
    success?
    
    Go read the article, tell your friends on Wall Street to look at it
    too.
    
    DJL

From:	PORI::MASTRANGELO "30-Oct-1992 1228" 30-OCT-1992 12:31:03.52
To:	LOUIS,RICK,DAVE,CARM,CHICK,RALPH
CC:	MASTRANGELO
Subj:	more on "NT on Wall Street (running on a Jensen)"

From:	MILPND::LJOHUB::NUSBAUM "Back in LJO2/C10, new DTN 226-2439  29-Oct-1992 1341" 29-OCT-1992 14:04:25.53
To:	countries,bpm,ipcstrng
CC:	
Subj:	Alpha and NT go to Wall Street -- with Bill Gates!

[forwards deleted]

From univax::bls Thu Oct 29 01:08:14 1992
Date: Thu, 29 Oct 92 01:08:06 -0500
From: univax::bls (Benn Schreiber)
To: ntsg-decwest, Jeff Schriesheim <jeff::jeff>,
        John Gilbert <decwet::gilbert>, David Stone <snkerz::stone>
Subject: Windows NT/Alpha demo'd at Windows on Wall Street
Status: R

Today at the Microsoft Windows on Wall Street Conference, Bill Gates
demostrated Windows NT running on Digital's Alpha AXP PC.  The stage was
set up with several monitors on the stage, with the systems hidden
behind the podium.  In full view of the audience of about 900 was a 16
processor Sequent MP system.

Bill started out his keynote with a discussion of the increasing
performance of microprocessors.  His first slide was a table showing the
relative performance of the 386, 486, and Pentium, along with the MIPS
R4000 and Digital's Alpha.  The table showed the initial chip
performance, and projected performance two years after product
introduction.  Digital's Alpha microprocessor was shown to have clear
leadership performance today, as well as in two years.

He then got a VERY big smile on his face, and did a demo of Windows NT
running on the Jensen system.  He demoed WINBEZ (the bezier curve demo),
the multi-thread demo, several aspects of the Program Manager, and
finished up with Solitare.

Bill commented during the demo that the port was ahead of schedule, and
that it was expected to ship on the Windows NT CDROM with the Intel and
MIPS versions.

It is especially worth noting that this was the ONLY demo that Bill
did during his one-hour keynote speech.  The original intention was to
demo an SQL server running on the Sequent, but it appeared that there
were last minute glitches, so it was not shown.  There were NO MIPS
systems on the stage, and only one MIPS R4000SC system to be found in
the demo area, and this system was staffed by an ISV (Canaan
Associates?) that had been working with MIPS/SGI.

There were actually two Sequent systems shown: the 16-processor
refrigerator-sized system shown on stage, and a smaller (roughly
Cobra-size) 6-processor system shown in the demo area.  These systems
use 486/50 processors on a VME bus architecture.  (I did not get much
information on the system architecture.)  However, they were running
the July PSDK (IDW 297), and the engineer I spoke with indicated that
they were experiencing some pain getting to the BETA system.  They run
the standard Intel binaries with their own HAL, drivers, and boot
code.  The port to this system was done by "about 8 engineers over 4-5
months".

Overall, I think that Digital's presence at this show was pretty good.
The DEC space in the exhibit area was kind of small, but we did have a
significant presence in the Microsoft booth, between the Jensen
running NT and the 486/50LP (Tiger II) systems.

   I attended "Windows on Wall Street" on October 28, 1992 representing my
   group, Capital Markets Practice.  Capital markets are the global markets for
   foreign exchange, stocks, bonds, etc.

   Microsoft organized "Windows on Wall Street" as a marketing event for their
   corporate customers in banking and capital markets.

   The overall impression is that Windows is here to stay, and that OS/2 and
   Unix are inconsequential.

   William Gates was the featured speaker and he introduced the current
   Microsoft strategy which is based on three core products: Windows, Windows
   for Workgroups, and Windows NT.

   In the description on Windows NT and in the Q and A, he gave a direct and
   enthusiastic endorsement of the Digital Equipment Alpha.  In the presentation
   he discussed Alpha as the primary RISC platform for NT and in the Q and A in
   response to question regarding what platform a user would be running NT on
   in the future, Gates said Alpha without hesitation.  The R4000 was mentioned
   mostly as an afterthought.

   What was new for me were the numbers: estimates that in its first 12 months,
   NT shipments will exceed that of all UNIX and all other large system
   operating systems: one million units.

   For the approximately 1000 people who attended, it was for the certainly the
   first time they saw NT on Alpha or perhaps NT itself.

   I was most impressed with Gates' style, much improved since the last time I
   saw him in person.  Indeed, he gave the clearest vision of computing in the
   future that I have ever heard and did it without the phony language of
   "strategic vision", "empowerment", "change process".

   In the sessions there was mostly a discussion of current work on Windows,
   the "developer councils".  There was one established for retail banking and
   one established for trading systems.  Some of the sessions were promoting
   vertical market products for the financial industry, others were
   customer-developed windows software projects.

   In the exhibit hall there was area of Digital hardware (Intel-based and
   Alpha) and my vertical market area demonstrating the latest Digital solution
   for trading, DECtrade for Desktop.

   The Message to Customers

   Windows, Windows for Workgroups, and Windows NT are the future of computing.
   Forget about OS/2 and UNIX.  For hardware, Microsoft runs on Intel and Alpha
   today and maybe other good RISC platforms in the future.  Microsoft is the
   safe bet. 

   The Message to Digital

   Ditto: Windows, Windows for Workgroups, and Windows NT are the future of
   computing.  I've made myself unpopular with friends at Digital who believe
   that UNIX is a safe bet and that Digital's investments in OSF are wise. The
   only safe investments for Digital are multi-platform investments, at least
   until the industry shakes out high-end desktop and server operating systems.

   Alpha's Windows NT availability is a key thing.  I suspect that some VMS
   sites will consider Alpha/OpenVMS to Alpha/Windows NT their migration plan
   from proprietary to open computing.

   Lastly, Microsoft's marketing, positioning, and execution of the Windows on
   Wall Street conference was superb.  The professionalism of how the event was
   conducted enhanced their message.

    re -1
    
    Pat suggests that some customers may consider VMS to NT migrations. 
    Since NT is very VMS-like, this sounds like it offers some real
    opportunities.  Has anyone heard of groups within DEC strategizing to
    do this?
    

    
    There are a number of organizations talking about this.  Hopefully
    things will be sorted out shortly.
    
    	Dave 

    I received a mail that contains the feedback from a focus group of
    Financial Execs. from wallstreet that summarizes their discussion on
    W/NT and Alpha.  It was done about a week before the Gates/Stone show,
    but it does shed light on how important this community sees our support
    of W/NT.  
    
    I don't think I can post it here because I don't know the originator,
    but I'll be glad to mail it out if anyone wants it.
    
    It echos the sentiments of some of the prior notes on the basis that
    the folks in the financial industry think NT will be Great and that if
    we really want Alpha to be accepted we must provide the highest level
    of support for NT.  This support should be given even if it is at the
    expense of VMS or OSF support.  They don't see VMS, OSF or OS2 as
    having long term futures in their markets.   The do believe that it
    will take a year of two for NT to mature.  
    
    They felt there is a need for an Alpha type Super-PC for the types of
    application that they run.  
    
    Their message was clear.  NT will make it, even if Alpha doesn't and
    that NT will allow them to have no loyality to HWD vendors.  Their
    loyality will lie with their applications.
    
    If you want a copy send an A1 to Furmanski@PTO or Haney::Furmanski

    .20 has been hidden until the author indicates he has permission to
    post the mail message in the conference.
    
    Bob - Co-moderator DIGITAL

(forwards deleted) - posted with permission of Kate Gardner
    
                  I N T E R O F F I C E   M E M O R A N D U M

                                        Date:     04-Nov-1992 04:21pm EDT
                                        From:     KATE GARDNER
                                                  GARDNER.KATE
                                        Dept:     Asset Management Marketing
                                        Tel No:   297-3807

TO:  Remote Addressee                     ( _fsoa::kgardner )


Subject: NT Focus Group                                              





   SUBJECT:  Windows NT Focus Group Findings - Detailed Report

   The purpose of this memo is to document detailed feedback
   from a Windows NT focus group conducted on October 21, 1992.
   The focus group was conducted by Digital with key New York
   banks.  The purpose of the focus group was to collect feedback
   on Windows NT and the Alpha PC strategy.  Digital's Windows NT
   and Alpha AXP PC strategies were presented by Michel Gambier and then 
   the floor was open for feedback and discussion.  The entire meeting
   lasted 4 1/2 hours.

   ATTENDEES
   ---------

   The twelve attendees were all senior IS executives from existing 
   Digital VMS accounts.  Their knowledge of Windows NT ranged 
   from non-existent to substantial. All of the attendees felt that 
   Windows NT would be very important in the future. They are ALL 
   planning to migrate from VMS to Windows NT over time; some earlier 
   than others. The group repeatedly said that Digital should make a 
   substantial commitment to Windows NT.

   The attendees included representative from  Participants Trust 
   Company;  The Bank of New York; Republic National Bank; Bank of America; 
   Bankers Trust; Advanced Systems Concept, Inc.; Chemical Bank; Barclays 
   Bank; Chase Manhattan; Pfizer, Inc.


   CURRENT PERCEPTION OF WINDOWS NT
   --------------------------------
 
   When asked about the key advantages and disadvantages of Windows
   NT, the group was quick to come up with advantages and had a harder
   time coming up with disadvantages.  They were asked about their
   perceptions, beliefs and concerns BEFORE they heard the strategy 
   presentations.

   o Advantages/Benefits of Windows NT

     - Windows NT will be a world class operating system in the PC
       space at the right price (they don't see this with VMS)

     - Windows NT is an operating system built with portability
       in mind

     - Scaleability

     - SMP
    
     - Client-server 
     
     - Virtual device drivers

     - Allows software to be disseminated to the desktop

     - Security

     - Better tools than OS/2

     - The ability to develop software for multiple platforms, cutting
       down on development costs - which is a significant benefit

     - Microsoft support is great relative to OS/2 and IBM

     - Believe there will be a variety of applications available on NT

     - Reliability		

   o Disadvantages of Windows NT

     - Further Windows NT release delays will allow other operating
       systems such as OS/2 to gain ground

     - Potential lack of applications (although the group felt that
       would most likely not be true)

     - Resources required to run Windows NT (part of the group 
       disagreed here, saying that they'd trade that for reliability)

     - It won't be ready for mission critical applications at Day 1
       since it's a new system and needs to mature before it's usable.
       It will need time to be accepted.

     - Migration might be an issue - looking for help here from vendors
       who will suggest solutions

   FEEDBACK ON DIGITAL'S WINDOWS NT AND ALPHA PC STRATEGIES
   --------------------------------------------------------

   o Digital's Opportunity With Windows NT/Digital's Commitment

   The group emphasized over and over again that Digital has a 
   tremendous opportunity to succeed with Windows NT and that 
   Digital should make a big commitment - even if it means shifting
   resources from OSF/1 and VMS.   Specifically, they said:

   - It doesn't matter what happens, Windows NT will succeed.
     DEC has no other choice than to go after it in a big way.
     Since DEC has limited resources, it's important to focus

   - Digital has an opportunity to go ahead and succeed with
     Windows NT; Digital has to get on the bandwagon full swing.
     Windows NT should be moved on aggressively.

   - If new resources can't be found (at DEC) to make a complete
     commitment to Windows NT, take them off of OSF/1 or VMS.
     UNIX is not as important as Windows NT.

   - Windows NT has more going for it politically than VMS

        - Digital must do something dramatic - DEC is hardly mentioned
     in the business press on Windows NT

        - Moving resources off of UNIX should be a no-brainer since
          Digital's UNIX perception isn't good, there's not much to
          lose

        - DEC should treat the move to Windows NT as serious

        - DEC has to hit big on day one (when Windows NT is released)

        - DEC has no other choice than to go after this in a big way

        - In order to be successful with Windows NT, Digital needs
          a Bill Gates equivalent reporting to Palmer; a corporate 
          champion

        - This transition (to Windows NT) is as important, if not
          more important as the transition from RSX to VMS

        - The Windows NT plans are the same as ALPHA VMS, so the
     group said they might as well make the move to Windows 
          NT now

        - The commodization of hardware should indicate that more
          needs to be invested in Windows NT by DEC

   - DEC should be the vendor of choice in PC-based servers
   
   - Current areas where Digital can add value to Windows NT
     include: TP, CASE, Services

   - There is a huge market for assisting in migration to 
     Windows NT

   o Digital's Service Strategy and Windows NT

     In general, the group was not satisfied with Digital's 
     service.  However, they indicated that there was a big
     opportunity in services for Digital - especially in
     migration.  For example:

   - Customer loyalty used to be to technology, but now
          they are loyal to a vendor because of their good
     service

   - Good service is always remembered - especially from
     a software company (i.e., Microsoft)

   - There is a huge market for migration services to
     Windows NT

   - DEC's consulting services have been poor; they'll
     go to smaller companies unless DEC builds up some
     core competencies

   - They will need help porting applications from VMS
          and would like Digital's help

   - If Digital doesn't allocate resources to port 
          applications, someone else will

   o Windows NT Product Strategy

     In general, the attendees had input in many product areas.
     However, their key inputs were in the areas of TP, DB, System
     Management, CASE (configuration management) and connectivity.  
     The group felt that Digital's software products should be 
     available VERY soon after Windows NT ships.  

     They also stated that enterprise system management tools, 
     diagnostic and performance tools were essential - these were 
     the tools UNIX never gave them but are needed.

     VMS Subsystem
     -------------
     The group asked if Digital was planning to build a VMS subsystem
     on top of Windows NT.  They indicated it should be done; but in
     the same breath said that DECmessageQ should be used instead since
     they'll all be using it.   They don't care too much about clustering
     since none of them use it today.

     TP
     --
     When asked about TP, the first thing that was asked was for
     TMX to be ported.   Everyone in the group felt that TP was
     critical to their success and was a big hole that Digital 
     could fill on Windows NT.  Specifically they said:

   - The TP monitor is essential - reliability is what keeps
     them awake at night

   - They would move to Windows NT because their maintenance
     costs are skyrocketing

        - Because it takes 2 years to build a TP application, they
     will begin working on their Windows NT port as soon as 
     they can

     DB
     --
     There was quite a bit of discussion about Digital's database
     strategy - most felt Rdb was going to have formidable competition
     on Windows NT from Sybase because they have a headstart with
     SQL Server.  Additional comments were:

   - Typical size of their database is 3-4 gigabytes

   - Digital should forget the database business on Windows
          NT

   - No one cares that Digital is moving Rdb to Windows NT

   - No one would buy Rdb on Windows NT, but they might buy
     	  a gateway   

     	- Ellipse is something they would consider in the future

     CASE
     ----
     The group felt that CASE on Windows NT was another opportunity
     for Digital, especially in the configuration management area.
     According to them, no one else was taking a global software
     development view and that Digital was one of the only capable
     vendors.  More comments:

   - No one is jumping on the Windows NT bandwagon for CASE;
     especially in the area of business modeling.  None of
     the major CASE vendors has asked any of the attendees
          to beta test their product on the Windows NT platform.

   - They'd like MicroFocus COBOL, IEF and Intersolv 

     Need for Super PCs
     ------------------
     Everyone in the room indicated they had a need for super
     PCs.  They'd like to look at multimedia, OO, neural nets
     and AI seriously but can't do it in today's environment.

     	- Natural data (voice, video, image) will play a big role
          in the 90s and they will need super PCs to support it

   - Software is the bottleneck now, software eats up the
     processor

     Power PC from IBM and Apple
     ---------------------------
     The attendees knew much more about Windows NT than they did
     about IBM and Apple's venture.  From what they knew about it,
     they thought it would be too late to market.  Comments were:
   
   - PINK might be interesting if it ever gets done

   - Windows NT is much more prominent - everyone senses the
     need to get on the platform

   - Lots of vendors have told the group they won't develop
     any more software for OS/2, they've been burned and
     will build for Windows NT instead

    Connectivity
    ------------
    When asked about connectivity needs, the attendees indicated they
    needed EVERYTHING so that fewer things would need to be changed.
    Mainframes need to be included.  File sharing and RPCs are also
    important

    Dumb Terminals
    --------------
    The group felt that there was a trend away from using dumb terminals
    because PCs are now cheaper.  	
 
    Applications
    ------------
    The trend is to buy sw applications rather than develop them.  All  
    would be very interested in a customer service application (maybe  
    something like the Client Management Framework?).  A specific 
    interest was expressed in trading applications, but there was also  
    interest in any type of banking specific sw.  Their loyalty is to the 
    applications, not the hardware or operating systems.  

   
    
    This document was written by Kate Gardner and Donna Slattery.


    .20 has been reposted with changes approved by the author of the mail
    message.
    
    Bob-Co-moderator DIGITAL

    Too bad we're only putting NT on one box.  A box that will wait till NT
    is ready.  If customers, think that NT will scale up and down the Alpha 
    line, they will be surprised.
    
    I was....but then, I was surprised to hear that HP is announcing the
    same day a 32bit box that has a higher specmark rating then our Alpha's.  
    And IBM and Sun are announcing that same day too!  Gee...
          
    

        It's a real shame that people are allready buying the media crapola
    that the HP machine will be faster than ALPHA. Knowing what some of
    those (ALPHAs) numbers really are I think it grossly premature to 
    assume the medai and HP are right. 
        Lets be patient and let the real benchmark data speak for Alpha and
    HP. If we can't have any (*(*^((* confidence in Digital why should our
    customers?
    
         Bob

    .25 is right.  Also remember the HP and SUN announcements
    are mostly program announcements.  The real systems won't be
    available for at least 6-12 months.  We are the only one that are
    ready to ship immediately.  Alpha is clearly the technology leader.
    
    This also shows how nervous our competitors are.  Although their
    systems are not ready yet, they still want to announce on Nov. 10th.
    Just want to show that they are not far behind (and the truth is they
    are).
    
    Wendy 
    
     

    Hey, a major customer benchmarked the two, HP and ALPHA and HP won the
    specmark test.  End of story. Not media crapola.

    ..did the customer end up paying for the minimal additional
    performance?

        Spiffy Keen Jim. There hasn't been a single revenue quality Alpha system
    released yet for any major customer benchmarking. None of the seed units
    of which over 1,000 were sent to major customers was capable of running
    full bore. As I said before media crapola. If you have hard data post it. 
    If not do DEC a favor and wait till next Tuesday.  We most certainly don't
    need any more negativeism from within. Besides if it's the alleged 
    benchmark I think your talking about HP is in for a "BIG" surprise. As an 
    aside all customers that were given seed ALPHA units were enjoined not to 
    perform any performance benchmarking. 
    
    
           Bob waiting for the big unveiling Nov 10th 

    re:27 sorry I've got your name wrong guess its not Jim....
    
          Bob

>>    .25 is right.  Also remember the HP and SUN announcements
>>    are mostly program announcements.  The real systems won't be
>>    available for at least 6-12 months.  We are the only one that are
>>    ready to ship immediately.  Alpha is clearly the technology leader.
    
     As much as all of us want to be optimistic, we can't be blind to
    realism.  We know pretty much what our performance numbers are right
    now.  There may be some small adjustments before initial shipments but
    they won't improve signficantly.
    
    It is has been widely speculated in the public press that HP's
    announcement will include a workstation on the order of 150 specmarks
    -- higher than ours.  I see no reason to doubt that they will do it
    based on their past history and the fact that they have done nothing to
    refute this.
    
    I have not see anything that indicates that the products they will
    announce are 6-12 months out.  Everything that I've seen indicates that
    they will be shippng these rather soon.  I may be wrong -- perhaps I
    just haven't see this -- but where have you seen evidence that this is
    nothing more than a program announcement?
    
    Remember that both HP and Sun ARE ALREADY attacking our "ready to ship
    immediately" message by the following:
    
    	1) OpenVMS AXP isn't up to the same level of functionality that we
           have on VAX/VMS.  Significant features (clusters, X-terminal
           support, etc.) aren't there at introduction and won't be for 1
           or even 2 more releases.
    
    	2) All layered products currently running on VAX/VMS won't be
           available for up to 2 years.
    
    	3) OSF/1 won't be shipping for at least several months. 
    
    
    We may believe that Alpha is the clear technology leader but it won't
    mean a thing unless our customers believe it.
    
    Be optimistic but be prepared.

Isn't it a shame that we didn't let Dave Cutler build his
latest operating system for us instead of Microsoft?

    re: .32
    
    > Isn't it a shame that we didn't let Dave Cutler build his
    > latest operating system for us instead of Microsoft?
    
    If Cutler had built it here, it would have sunk into obscurity.
    The magic words for NT are "Microsoft", "Gates", and "Marketing",
    not "technology" and "Cutler".
    
    Most of what I've heard from our customers and some of my high-tech
    friends in the area bear out the bankers' emphasis on NT, to wit:
    
    - The demand for a shrink-wrapped O/S with scaleability and portability
    can't be underestimated.  UNIX has failed to live up to a lot of
    people's expectations in this area, and NT is the new hope.
    
    - Digital's lukewarm support of UNIX has created customer fears that
    we will *never* fully support anything NIH (not invented here).  We
    must break these perceptions, or we can abandon the lion's share of
    the future market.
    
    - Microsoft and Bill Gates are perceived as winners, while our own
    fortunes have dwindled.  It's not a technical evaluation, but it
    *does* influence buyers.
    
    - We are finally getting some good press about our PC's, but a lot of
    customers are still concerned that the Alpha is a trojan horse to get
    everyone back into the proprietary minicomputer dungeon.  Windows NT
    could help us dispel this concern (it *is* unfounded, isn't it?).
    
    Geoff Unland in Austin

    I'm not going to announce who the customer is, since they are in legal
    trouble from both HP and DEC for disclosing the benchmark test results.
    
    But, if you think they have not done the tests, then you are wrong or
    misinformed. They have and we lost the Specmark test. We did win some
    tests though...

    
    Quit arguing about which is faster. The real message is:
    
    	"Alpha is not alone; it has competition from day one, and 
    	 it only gets tougher from there." 
    
    The days when the speed differences in chips from different 
    companies was dramatic are over. 
    

    Has anyone noticed that WNT (for WINDOWS NT) is VMS + 1 (as in V+1, M+1,
    S+1)?
    
    Similar to how Arther C. Clarke named the HAL 9000 computer in "2001: A
    Space Oyssey" (IBM -1, as in I-1, B-1, M-1)
    
    One of my co-workers pointed this out to me this past week.
    
    Anyone want to speculate if this thing with the initials was intentional 
    on Cutler's part, or "merely coincidental"?  I think it was planned.
    
    
    
    Bill Slater

    Re: .36
    
    This circulated several months ago - supposedly Cutler, when asked
    about it, said "What took you so long?"  However, I don't think
    it was deliberate and I don't share some people's lofty opinion
    of Cutler's work.  I just hope for all our sakes that NT is a lot
    better out of the gate than any of the software Cutler produced for
    DEC.  (And if history tells us anything, Cutler will likely abandon NT
    as soon as it is released to go look for something more "interesting"
    to do, leaving others to clean up afterwards.)  
    
    			Steve

>    I'm not going to announce who the customer is, since they are in legal
>    trouble from both HP and DEC for disclosing the benchmark test results.
 
    And well they shoulf be in legal trouble. I hope we and HP make a
    serious example of them.
       
>    But, if you think they have not done the tests, then you are wrong or
>    misinformed. They have and we lost the Specmark test. We did win some
>    tests though...
    
    I don't think that anyone doubts that tests have been run. Just that
    the tests are reasonable given that full speed Alpha H/W and S/W isn't
    even available within Digital yet. Or at least not in any quantity. And
    no customer has full speed Alpha AXP systems.
    
    			Alfred

    RE .37
    
    Steve, I've read several notes in various places written by you that
    belittle Dave Cutler. But you never give specifics. I think if you are
    going to trash a person you should at least explain exactly what you
    base that opinion on.
    
    I know you come from the compiler space. I have more often heard the
    compilers that Cutler did (VAX C and PL1 I believe) criticized for
    quality than his OS work (RSX and VMS). Could that be because OSs
    are his real interest and compilers was something he just marked time
    with?
    
    The way I look at it is that RSX-11M and VMS have made this company a
    lot of money. I believe it is certainly a fact that Dave Cutler had a
    significant part in both.
    
    The above said I'm really curious on what facts you base your opinion.
    
    Re .-1
    
    Why is HP upset with this customer? I thought the results put HP in a
    really good light.
    
    Dave

    I would assume that HP is glad that some results make them look good
    but they'd probablly rather have their performance numbers kept under
    wraps until *they* are ready to release them. Timing of announcements
    is a big deal to companies. One doesn't want the other people to know
    what target to shoot at early.
    
    			Alfred

    Sorry this is little of-line on the Alpha chip talk, but about Cutler, could
    some one tell us please something about his background? I've heard
    his name a number of times, but know little about his background, like
    what did he do before coming to DEC? did he come from academic to 
    industry, is he a PhD in computer science? (not that it makes a 
    difference, just curious) what was his title when he was in DEC ? I 
    bet he was at least a principle software engineer or even a consultant
    (from what I hear he worked on)..
    
    thank you very much,
    /nasser

    Yeah, VAX C was not one of the Cutler success stories.
    Nor was VMS DCL, now I think about it.
    
    Cutler seems to be really pretty good on squeezing the last
    cycle out of a CPU.   I also don't think you can fault him on
    his ability to bring things to market (I don't mean to
    imply by that that he's unassisted) and from what I
    can tell he's a prodigious coder.   You probably can
    quibble with him on general elegance, though.  On
    the other hand, I liked what I heard about NT.  It
    certainly sounded elegant enough the way he described
    it.   

    HP is upset cause while we lost the SPecmark test, when they rean their 
    application, it ran 13 times faster on ALPHA.
    
    

    re: .32  With respect to "why didn't we do it here," Digital doesn't
    really have a history of embracing new technologies that would compete
    with its "bottom line".  (U*ix comes to mind.)  Also, NT is out in test 
    form now and it is like pulling teeth to get some departments to spring 
    for a platform to do NT development on.  In our case, it's because our 
    (internal to Digital) customers don't currently demand NT and are expected 
    to use VMS only.  Yes, they are turning to outside vendors for support and 
    are starting to use PCs, but the "real work" will all be done on VMS 
    machines. That's what I'm told.  So, I (like many others) am building my own
    NT-compatible system *at home* and plan to do NT development and
    get application porting experience and support there.  From what I've
    seen, Digital is NOT (as a company) trying to take any sort of leadership
    position as far as NT is concerned.  Rather, it is waiting for our
    customers to demand it and maybe place some orders before we really
    commit to it.  We've seen it all before.  It's what happened with U*ix.
    
    re: .35  I think the point of Alpha is that it is an architecture that,
    unlike many others (X86 family comes to mind), was designed for growth.
    Some folks think of Pentium as being a direct competitor to Alpha. 
    But, keep in mind that while Pentium may well be at the high end of its
    family, I understand that Alpha is the low end of its family.  It's the
    architecture, not the raw speed, that impresses me as far as Alpha is
    concerned.  Speed is not the only issue.  The other issue is
    architecture and ability to grow with the technology.  While others are
    hoping to match the performance of Alpha, my understanding is that we
    are already working to beat our own best efforts with an architecture
    that should support advances in technology and cost effectiveness.  
    
    Steve

    Re: 2186.44
    
    Steve,
    
    You are a little off the mark in assuming that "Digital as a company is not
    providing any leadership" in planning migrations to WINDOWS NT on AXP
    machines.  I work in the Worldwide ALPHA Migration Center in Merrimack,
    NH, and I can tell you that we are actively working on strategies to
    migrate from several different source environments (DOS, WINDOWS 3.1,
    OSF/1, and U*IX) into this WINDOWS NT target environment).  
    
    I don't want to drag this note into a rathole, by discussing migration
    strategies here, but rest assured that Digital is pursuing this type of
    migration and several others for our customers who will be purchasing
    ALPHA.
    
    BTW, I used to work in Houston with a Steve Sherman.  That was for a
    large energy company.  You aren't the same Steve I worked with are you?
    
    Regards,
    
    
    Bill Slater
    DTN 264-4953
    
    same one

    re: .45
    
    I am aware that segments of Digital are very committed to NT.  But, I
    still feel that Digital as a company is not providing the leadership.
    Were I wrong, I would expect that NT would be more widely accepted and
    supported within Digital as are OSF standards.  But, I and others are
    encountering resistance to NT within our groups.  We're having to
    resort to building our own NT machines at home if we are to get our
    hands on the technology at all.  What company that is establishing itself 
    as a leader in a technology puts their employees in this kind of situation?
    
    Steve

    Steve, sounds like you report thru the wrong stovepipe. :-)
    
    Seriously, Digital will probably continue to do other things so maybe
    it isn't necessary for everyone to follow the same strategy.  An old
    friend of mine used to criticize Digital because the company seemed to
    try to move everyone in the same direction at once instead of
    recognizing that we need choreagraphy more than the chorus line.
    
    Dick

    Hi, Dick!
    
    Actually, I'm really not grumbling as loud as it might seem.  It 
    *IS* kind of nice to be able to (finally) be able to get a 
    development system at home.  And, in a way it brings a bit of 
    comfort because it's *my* system, with all that that implies.
    
    One benefit is that of being my own system manager.  I don't
    have to wait to install the latest software.  And, the software
    is cheap albeit not as reliable as VMS stuff (as a rule).
    But, I can get real work done on it.  I can do things on it
    that I can't do on the machines at work and the results are
    professional-quality.  I like the workstation at home for the
    same reasons that I fell in love with the VAX nearly 15
    years ago.
    
    When I learned about NT, CD-ROMs and multimedia applications,
    I was at first shocked by how much a PC can do that I can't
    do with my workstation at work -- not because it can't, but
    because the funding is just not there and the expense would
    be to high to justify it in terms of Digital's survival.  
    The irony is that for a home machine the costs *CAN* be 
    justified in terms of survival of MY career, especially during
    uncertain times because NT capability will likely increase my 
    ability to remain gainfully employed -- inside or outside of 
    Digital.
    
    Steve

RE: .32

Dave Cutler DID start building his latest operating system at DEC.  It was 
called MICA and was to be the operating system for the PRISM RISC processor.  
It was to be a highly portable, microkernel operating system with VMS- and 
Unix-compatible API subsystems running on top of it.  Both PRISM and MICA got 
jerked around and redefined several times due to politics, indecision, and 
strategic waffling at the top levels of the company.  Eventually both projects 
got cancelled, which was just as well because by that time they'd been 
compromised to the point where death was probably the best thing for them.

Dave Cutler left DEC shortly after the PRISM cancellation to work on a highly 
portable, microkernel operating system with several API subsystems running on 
top of it.  We know that OS today as NT.  The NT/Alpha AXP combination 
resembles quite strikingly the original MICA/PRISM concept.

--PSW

RE: .39 (What is it that DC did poorly?)

Steve Lionel speaks from the experience of observing what happens for several 
years after Dave Cutler leaves one project to start another one.  I can speak 
from the same experience.

What Dave is best at is doing the planning, management, design, and 
implementation of high-risk, version 1 products.  If I had such a project and 
it was critical that it get out the door on time, no matter what, Dave Cutler 
is the person I'd choose to lead it.

What DC has historically not been particularly good at is engineering for the 
long term, especially engineering for maintainability.  He has had a tendency 
to ignore past experience from similar projects (VMS V1 used none of the good 
ideas present in TOPS-10 and TOPS-20 because DC didn't get along with the 
people working on those products).

DC has done some excellent work for which he justifiably deserves the credit.  
However, I've also seen his name on some god-awful, unmaintainable messes that 
somebody at his job skill level (Senior Corporate Consulting Engineer) has no 
excuse for perpetrating.

At DEC, Dave Cutler did not walk on water.  Rather, he dredged out the river, 
and tended to leave piles of reeking muck lying on the banks.

--PSW

According to Ed Yourdon in "Decline & Fall of the American Programmer"
it is never exusable to ship on time with less than high quality.
Can an engineer good at hacking it up fast fit into modern software
industry?  If they go to a company famous for shipping low quality on time,
is that a match made for America's future or for Japan's?
Tom

But the question is not shipping a quality product, but how much quality
product should the product contain?

A 100% defect-free product is definitely high in quality, but it probably cost
prohibited.  A 98% defect-free product might seems to a customer to be the same
as 100%.  Also, if you have *excellent* customer service, then a slightly 
defective product might be better.  Why?  If the 100% product never fails, the
customer will take it for granted and may forget about (out of sight, out of
mind).  Now the slightly more buggy product (cosmetic bugs only) and appropriate
stroking by customer service, the customer will remember how great X is and
spread that via word of mouth in most instances.  Now that could lead to more
sales and ...

The quality of product is not an absolute measure, it is something that will
traded againt time to market -- what needs to done is to find an appropriate
balance.

>            <<< Note 2186.52 by NETRIX::thomas "The Code Warrior" >>>
>
>But the question is not shipping a quality product, but how much quality
>product should the product contain?
>
>A 100% defect-free product is definitely high in quality, but it probably cost
>prohibited.  A 98% defect-free product might seems to a customer to be the same
>as 100%.  
One bug in 50 keystrokes? I don't think so.  98% or 99% have been unacceptable
levels of quality everywhere for years.  Especially fixing it at our cost
under warranty.  We bear the cost of fixing things, and fixing bugs costs
orders of magnitude more to us than making it correctly in the first place.
>Also, if you have *excellent* customer service, then a slightly 
>defective product might be better.  Why?  If the 100% product never fails, the
>customer will take it for granted and may forget about (out of sight, out of
>mind).  Now the slightly more buggy product (cosmetic bugs only) and appropriate
>stroking by customer service, the customer will remember how great X is and
>spread that via word of mouth in most instances.  Now that could lead to more
>sales and ...
It's the opposite.  Customers remember the bugs and buy elsewhere.
>
>The quality of product is not an absolute measure, it is something that will
>traded againt time to market -- what needs to done is to find an appropriate
>balance.
No.  Quality must never be sacrificed for time to market.  Ever.  The major
bug in Digital's process is the addition of excessive numbers of features.
It becomes easier to avoid bugs AND make time-to-market by not allowing the
project management and engineers to invent features that are not justified
in the customer interviews or the market and competitive analysis.

"Quality is first!  *Always* first!  If we deliver bad quality to the customer,
the customer will complain many times, over and over.  But if we are late, 
he will complain only once."
Tomoo Matsubara of Hitachi Data Systems, quoted in Ed Yourdon, The Decline and
Fall of the American Programmer, Yourdon Press Computing Series  1992

-Tom

I'd have to go back and find out where I read it but what I described took place
at Polaroid.  They came out with a line of cameras which basicly were defect
free.  They sold well enough but there was only a 50% repurchase rate of the
customer.  They then came out with a line that a slightly higher defect rate
but had an 800 number and the cusomter staff was responive, sent flim, flashes,
etc.  That line had a 85% repurchase rate.

Quality is a subjective term -- it has no absolute measure.  What I feel is a
quality product you may feel is a bug-ridden product.  Who's right?  We both
are.

A product should have no major bugs, period.  But about less obvious bugs?
That will never happen in 1 out of a 1000 sites?  Do you hold up the product
for 2 months while you fix it?  I think not.  You release note the problem and
ship it *and* follow on with a *fix* ASAP.

One of our major problems is that it takes too long to get something fixed once
a customer reports a problem.  Recently I was involved in two different
problems.

One was a data corruption blamed on DECnet on a specific platform.  It took
about 6 months to reach my group but once we got and reproduced, we identified
an ULTRIX driver bug and generated and tested a fix.  Once the problem was
understood it took about 2 hours to find the bug in the source and develop
a fix.

The another was a bug which had been kicking around the corporation for 4 months
before the customer got frustrated enough to post something to USENET.  I posted
an incorrect answer, alas.  The customer finally had a name and got his local
office to track me down.  Once I got the problem report, it took two minutes
to fix it and in addition we added another feature to something he was to make
his life even better.

If the above are typical of our bug-fixing procedures, then we must ship
defect-free products since we will never get them fixed after they ship.

Cisco (who is kicking our ass in the router market) has never been accused of
selling high quality software (it works but it had quirks), but they do provide
is very fast and very accurate bug fixes and tech support.  Much of their tech
support is done via mail over the Internet (check out comp.dcom.sys.cisco).

    And don't forget Microsoft. They have a reputation for shipping buggy
    V1.0 products, but they still sell.
    
    Ken....
    

I saw the following in a comp.software-eng posting and it seems to
apply to us. It also gives a customer's perspective on finding bugs
and customer service.


> One of my pet peeves is that certain companies who will not work on a bug
> unless it is submitted through a cumbersome procedure which guarantees
> that only critical bugs for customers who buy support get worked on.
> I think this is extremely shortsighted.  Any company should be grateful
> that anyone, customer or not, has reported a bug, and all serious bugs
> should get fixed ASAP.  In effect, someone has done some *free* testing,
> and the company does not utilize the result.  If they aren't going 
> to accept free testing, you can't expect them to *pay* someone to do it,
> can you?

    I can recall a conversation I had with an IBM buddy who called their
    practice of shipping 'something' on time (to be later define/fixed by
    customer feedback) as Market Driven Quality (what he said internally
    was known as 'just enough engineering')
    
    this one reason NT will be successful is that they'll get it out the
    door and let the customer tell them/pay to fix it; I'm helping an
    account now on a PC client/server system, where the customer is paying
    MicroSoft $2500/day to dianose a problem in a himem.sys driver on a NEC
    pc; if that was our equip. you'd know we'd be giving it away
    
    /los

re .37, .39, .41, .42 & .50:

     You can pick this discussion up in FIGS::CUTLER$BRICKBATS$AND$KUDOS if you
want to.

   Sorry for reopening old wounds, but reading this I couldn't help but
respond...

>         <<< Note 2186.55 by NODEX::ADEY "I don't do Windows...NOT!" >>>
>
>    And don't forget Microsoft. They have a reputation for shipping buggy
>    V1.0 products, but they still sell.
>    
>    Ken....
>    

   Microsoft's V1 product quality exceeds the quality of many DEC V2 or later
products, IMHO.  [And I'm not claiming that MS's are high quality!]

   They have to - it just doesn't make economic sense when you have to
manufacture hundreds of thousands of units of shrink-wrapped software.

   I personally usually can live with it [DEC software quality] because I have
more-or-less direct access to developers, bug-fixes and early field tests of
the next version. When I consider what it would be like to use our software
ourside the company, I cringe and wonder...

					-mjg

    re .59
    
    Ah, savage product memories!
    
    You brought to my mind DECsystem-10 COBOL V10, which had the dubious
    honor of requiring a 10A release because nobody had anticipated a
    product whose patch counter field (what was the real name - it's been
    quite a while) would need more than 3 digits.
    
    DBMS-10 V1 was also a joy.  The manual was a work of pure fantasy, and
    the schema compiler was, well, eccentric.  To say the least.
    
    Hmmm.  Better stop before I wax contemporary...
    
    /Butch

Re .60:

>    You brought to my mind DECsystem-10 COBOL V10, which had the dubious
>    honor of requiring a 10A release because nobody had anticipated a
>    product whose patch counter field (what was the real name - it's been
>    quite a while) would need more than 3 digits.

Must have been some self-inflicted internal field - plenty of room for quality
improvement in the system-defined version layout for MAJMIN(EDIT)-WHO:

SUBTTL	INTERESTING SYMBOLS

;VERSION FORMAT COMPONENTS

VR.WHO==7B2	;WHO EDITTED (0=DEC DEVELOPMENT, 1=OTHER DEC,
		;     2-4=CUSTOMER, 5-7=END USER)
VR.VER==777B11	;MAJOR DEC VERSION
VR.MIN==77B17	;MINOR DEC VERSION (1=A, ETC.)
VR.EDT==777777	;EDIT LEVEL
				/AHM

VR.WHO==7B2	;WHO EDITTED (0=DEC DEVELOPMENT, 1=OTHER DEC,

Thus proving that it's not only 32-bit programmers who can't spell.

>Thus proving that it's not only 32-bit programmers who can't spell.

    thanks Dave for pointing this out, i been trying to expalin this all
    along but no one believed me.

    \nasser
    a_bad_speller_becuase_iam_a_32_bit_programmer 
    

Re .62:

No, they spelled it right, at the time the code was written the author
undoubtedly considered that they worked for DEC.
				/AHM

    Re .-1
    
    It's not DEC that is incorrectly spelled.
    
    Dave

    Dave Garrod's been in the USA far too long - he can no longer
    recognise a joke when he sees one (unless, of course, it has
    "ha, ha, only kidding" written on it).
    
    Only kidding, Dave  :-)