Conference 7.286::digital

    Sounds bitter to me.  Also seems unaware of FX!32.
    
    

    
         "Thought provoking" ??  I disagree.  Nothing was said that hasn't
    been said dozens of times in this note conference.  Beyond that, it had
    the heavy flavor of a severely disgruntled ex-employee.
    
         This is someone venting their frustration.  That does not mean the
    sentiments are not valid.  It just means that it added nothing to
    cosmic consciousness evaluating the future of this corporation.
    
                                     Greg

    Not Just Bitter...
    One of the 60,000 Landmines Sales and Sales support has to deal with 
    every day we try to make money for Digital...
    
    JMHO..
    
    John W.
    
    

re.0:

>    But they do have Alpha, the world's fastest CPU, if only people did
>    not want to run software, but I am sure it is only a matter of years
>    before they have a fraction of the applications of other vendors.

Is it possible that the customer has some special requirement or that they
just have the wrong impression of the status of applications ported?
Perhaps you should reply and get this point clarified by asking what
applications they would like to see on Alpha? It may be that the 
applications are amongst the 8000+ already ported.

Dave.

  .5  can he say that in here?

    
    <--- with impunity no less.

> Is it possible that the customer has some special requirement or that they
> just have the wrong impression of the status of applications ported?
> Perhaps you should reply and get this point clarified by asking what
> applications they would like to see on Alpha? It may be that the 
> applications are amongst the 8000+ already ported.

While it's getting better, there are still a LOT of cases where
apps either aren't available, were available but were withdrawn
due to lack of volume, or lag the competition significantly (arguably
also due to lack of volume.)

So while quoting numbers like "8000" may seem impressive, it doesn't
help if they aren't the apps you're interested in.  (Examples
of some things - Netscape has lagged on Alpha NT a lot, though
it's getting better, a set of TCP products including an NFS server
was dropped due to lack of volume, Office 95 isn't available native
for Alpha NT, etc.)

In the current PC Week there's a letter to the editor blasting Digital for
having "suits who don't care".  The letter is on target, and I suspect was
written by one of the tens of thousands given the ax.

				Steve

There is a big difference between whether an application can be made to run 
on an Alpha (via FX!32, for example), and whether the application vendor 
will support the translated application.

And the issue of drivers and their use in applications is even more 
important.

For example, suppose there is a 32-bit OCR package that runs on Intel/NT 
using some special scanner.  First, you have to have an Alpha driver for a 
given scanner.  Second, if there is any problem running the application, 
will the vendor give me "click, buzz" support when I tell them their 
Intel/NT app has been translated to run on Alpha.

Similarly, a video conferenceing package will need not only FX!32 
translation for the app, but a driver for the camera.  Intel/NT drivers 
for specialized hardware are tough enough.  Alpha/NT drivers are a true 
rarity.

Clearly Alpha is in a much better position than MIPS or PPC, but it will 
still raise application vendor support issues for customers.

john

re.8:

>While it's getting better, there are still a LOT of cases where
>apps either aren't available, were available but were withdrawn
>due to lack of volume, or lag the competition significantly (arguably
>also due to lack of volume.)

True but we can't be all things to all people. We need to focus on specific 
industry and application areas to provide the ISVs with the volume they 
need. This is a long term strategy that does not fit well with a company 
that reorganises itself every quarter. 

Dave.

> True but we can't be all things to all people.

But what you're really saying is we can't even take on one of the 3 markets
that we're claiming to be going after.

> This is a long term strategy that does not fit well with a company 
> that reorganises itself every quarter. 

Launching a new architecture is a long term strategy, and if the company
can't stick to the things necessary to make it succeed in a very competitive
marketplace, it WILL fail.  From my vantage point, it certainly looks like
we're stuck on the short end of a feedback loop that can either be positive
or negative.

re.12:

>But what you're really saying is we can't even take on one of the 3 markets
>that we're claiming to be going after.

No, I'm not saying that. I'm saying we have to attack those markets in a focused
way in order to be successful. 

Dave.

    
      how bout a new Alpha Applications catalog.  
    
     last one I've seen is dated Nov'94 !
                                                

It's on the web - http://www.unix.digital.com/catalog/index.html
(don't get confused by the "unix" - this is for UNIX, VMS and NT).

					Steve

    .-1 is ANOTHER example of:
     
     - Everybody gets it all from the WEB (go to our RESELLERS offices,
       MANY still use dumb terminals with no thoughts of PCs) they
       simply order fulfill and do NOT do demand generation
    
    And in general the comments from Digital folks about "sounds
    disgruntled"....well..simply go to your computer store or reseller and
    ask about OUR products versus a competitive one. Take a stroll and
    you'll be shocked.  Doing it better does NOT matter...doing what the
    customer WANTS quickly and responsively is the ONLY DAMN THING that
    matters!
    
    Finally, the comment about "thousands of land mines" in the field every
    day could not be more true.  The average Digital corporate person's
    life expectancy in the field is about 4 weeks.  (Remember our "career
    opportunity days"?  Nearly all of them toasted in the first year.  They
    could not believe what the field had to deal with.  And if THEY could
    not deal with it, our reseller surely could not..)
    
    

>  <<< Note 4728.10 by ANGST::tun-4.imc.das.dec.com::boebinger "John Boebinger - (330) 863-0456" >>>
>                              -< support means more than translation >-
>
>There is a big difference between whether an application can be made to run 
>on an Alpha (via FX!32, for example), and whether the application vendor 
>will support the translated application.
>...
>For example, suppose there is a 32-bit OCR package that runs on Intel/NT 
>using some special scanner.  First, you have to have an Alpha driver for a 
>given scanner.  Second, if there is any problem running the application, 
>will the vendor give me "click, buzz" support when I tell them their 
>Intel/NT app has been translated to run on Alpha.
    
No John, there really is very little difference.

What about 16-bit Windows code that runs under the Windows NT VDM
emulator? Do these software vendors not support their code for this
"virtual" platform???

Java code is entirely emulated. Do Java vendors not provide support???

Furthermore, our FX!32 group is ready and willing to provide support
for any problems in running Intel NT applications under Alpha NT.
Call me personally if you have problems.

George Darcy
Alpha Migration Tools
DTN 227-4109

	  I'm not John, but ...

       >  What about 16-bit Windows code that runs under the Windows NT VDM
       >  emulator?  Do these software vendors not support their code for
       >  this "virtual" platform???

     That is correct -- many vendors do not support their code on "virtual"
     Windows platforms such as NT and OS/2.  I have had commercial
     applications for Win 3.1 not work on Windows NT (Intel), called the
     manufacturer, and was told that was not supported.  I protested
     Microsoft said it would work -- the vendor told me, not unreasonably,
     to call Microsoft.

				     Brian

     Brian,

     Digital plans on *fully* supporting FX!32. In the unforeseen circum-
     stance where your software vendor cannot provide support due to system
     type, our group will get involved to ensure that your Intel NT applications
     runs the same on Alpha NT. We have 30 people here testing hundreds
     and hundreds of applications to ensure our success.

     The hardware/software distinction is bluring anyhow. Java programs
     run completely under emulation (interpretation and translation).
     Also, Microsoft is also working on its own RISC based interpreter
     for NT. There wouldn't be all this activity in emulation, if people
     did not have faith in the technology...

     Regards,
     George

  Well, it can't be an "unforeseen" circumstance; Brian just told
  you it's a fact of life, even on an Intel-to-Intel emulation.

  Secondly, why would I (as a customer) want to select an emulated
  solution (which depends on a third-party whose support is now get-
  ting trashed in public) when a native solution provides equal or
  nearly-as-good performance, and *WON'T* cause me any hassles get-
  ting support from the original software vendor?


> We have 30 people here testing hundreds and hundreds of applications
> to ensure our success.

  TODAY you have 30 people.

                                      Atlant

  >Secondly, why would I (as a customer) want to select an emulated
  >solution (which depends on a third-party whose support is now get-
  >ting trashed in public) when a native solution provides equal or
  >nearly-as-good performance, and *WON'T* cause me any hassles get-
  >ting support from the original software vendor?
    
    If there is a native Alpha version of the application, then
    certainly you would use it.
    
    People buy Alpha for its high performance in running some
    set of native applications. FX!32 then completes the story by
    telling the customers not to worry about those other apps for
    which you do not have "Alpha" bits. FX!32 will give you good
    performance, comparable to Intel, for those applications.
    
    Rgds,
    George

    I think Atlant, in .21, meant 'native' as in 'native on an Intel
    platform' ..
    
    ...tom

  >TODAY you have 30 people.
  >
  >                                    Atlant
    
    Yes we do. Are you surprised?
    And our group is growing.
    
    George

re: .23

>  >TODAY you have 30 people.
>  >
>  >                                    Atlant
>    
>    Yes we do. Are you surprised?
>    And our group is growing.
>    
>    George

I believe Atlant's point is that, although you may have 30 people today, 
given events of the past few years, there is no way of knowing how many 
people you will have in six months, or whether you will even be in business 
in six months.

Ray 

>I believe Atlant's point is that, although you may have 30 people today, 
>given events of the past few years, there is no way of knowing how many 
>people you will have in six months, or whether you will even be in business 
>in six months.
>
>Ray 
    
    No one has a crystal ball. But if our group's past history is any
    indication of future success (like in mutual funds), I would
    say we are on track for success.
    
    Regards,
    George

    The last part of Digital to go under will be the Semiconductor group.
    The most important thing in the Semiconductor Group's strategy for
    the foreseeable future is FX!32.
    
    In other words, these 30 people have the safest, securest job in the
    company.
    
    Frankly, I think the FX!32 strategy will be blindsided by $2500 PPro
    200's, but it'll take 'em at least a year to figure out why that
    happened.
    .02 Kratz

The issue is one of risk/benefit.  There are a fair number of coporations 
who believe that the right thing to do is to standardize on Windows NT for 
the desktop once NT V4.0 ships.  And they have very good reasons for doing 
so (stability, etc).  This would seem to be a golden opportunity for Alpha.

However, assume that you are the CIO of a large New York bank.  You plan on 
rolling out 10,000 desktops running NT in the next few years.  Your choices 
have been narrowed to Alpha or Pentium Pro.

The Alpha will give you higher performance.  But it is in the range of 50%, 
not 2x or 3x.

With the Pentium Pro, your software packages (OCR, FAX, financial analysis, 
etc) will be directly supported by the vendor.  This make support 
straightforward.  Internal applications can be written using Visual Basic 
V4.0 fully supported in 32-bit mode.

With the Alpha, some (Microsoft Office) may be directly supported, others 
will require that Digital support them through the 30 people in the FX!32 
group.  You have to trust that Digital will continue to maintain that level 
of software support for each and every one of the applications you plan on 
rolling out.  VB V4.0 is not fully supported on Alpha.  Maybe V5.0 will, 
but no one's talking.

If there are major problems with supporting software based on your choice 
of hardware platform, you discover that there is such a thing as being 
fired for poor performance.

Which platform do you choose.

Please recognize that this is not an issue of technical excellence.  At any 
given point in time the Alpha is faster than Intel.  We've done this with 
much less engineering investment than Intel.  FX!32 is a great piece of 
code that deserves much praise.

But these factors are not relevant to the issue of which platform will most 
likely allow the CIO to remain employed.

john

>Also, Microsoft is also working on its own RISC based interpreter
>for NT. There wouldn't be all this activity in emulation, if people

Hunh? How does this affect/relate to/compete with/complement FX!32?

	C'Ya,
	Wayne

  Both folks interpreting my words interpreted them correctly --
  Thanks!

  Yes, I meant native on Intel. That's the big market and you
  can bet that's where the vendors will concentrate their
  efforts. They won't be concentrating on "boutique" RISC
  engines even if they deign to support them a little bit
  while Intel catches up in performance.

  And yes, I absolutely meant that, based on very recent
  history, your group could evaporate at any moment, NO MATTER
  HOW WELL IT'S DOING or HOW STRATEGICALLY IMPORTANT IT IS.
  That's just a fact of life here at Digital today. Yes, you've
  got 30 people ready, willing, and able to help me today, but
  if I bet my company on your solution, will you be there
  tomorrow? Maybe I'd better just bet on native Wintel instead.

                                     Atlant

  >Well, it can't be an "unforeseen" circumstance; Brian just told
  >you it's a fact of life, even on an Intel-to-Intel emulation.
    
    From experience I'd say that about 1/4 (or some sizable fraction)
    of all your app installers on NT are all or part 16-bit,
    and thus execute directly under the VDM. So emulation is not
    some quirky, odd-supported mechanism for installing and running
    applications. It's been proven. And Microsoft wouldn't be spending
    money on its 32-bit emulator if it thought otherwise. Where are
    all the cries against Java? Can you say "Code Byte"?
    
    George

George:

 You're missing the main point.

  Let's suppose you're a customer, and you're trying to run
  ZumaSoft's new "SurfWriter" product ("the word processor
  that's, like, totally gnarly").

  On a Win95 platform, it's running native. That's also 95%
  of ZumaSoft's market. On the Alpha, it's running under an
  emulator (and is, say, 1% of the market). And, BTW, their
  equivalent of an SPD doesn't mention Alpha, RISC, or emu-
  lators, but it *DOES* mention x86, Pentium, and Pentium
  Pro.

  Well, one day, there you are, cranking through your words
  on your Digital Alpha and suddenly it's like "Bummer! The
  SurfWriter's gone belly up, it's a former word-processor,
  it's like totally toasted!"

  How excited will ZumaSoft be to get your phone call, *ESPECIALLY*
  if they can't reproduce this wipeout on their Wintel machines?
  To whom will you turn next? Will they still be there to support
  you and your emulator? *THAT'S* the issue that several of us
  are trying to get you to see, not the issue of whether FX!32
  or VDMs are feasible in the first place. We'll grant you that one.

 -=-=-=-=-=-=-=-

  BTW, you keep mentioning "Java". This question (of code developed
  to run on one architecture and ported to others via emulation)
  really has very little to do with Java (which is pseudo-code
  *DESIGNED* to be architecture independent and *INTERPRETED*
  by interpreters running on various architectures). With Java,
  architectural portability was a main design goal. It most
  certainly wasn't/isn't in your average Wintel app.

                                       Atlant

    Who cares what bits anybody twiddles to do what.  If my job is on the
    line I want it to work, do it quickly, easily and help me run my
    business. 
    
    We need to remember that all the technical gee-gaws that we make,
    Alfers included, are just ARTIFACTS of the customer's real business.
    
    Get it.
    
    We only make ARTIFACTS that help other people do their job.
    
    Our stuff needs to be simple, clear, widely accepted, easy to use. 
    
    So stop arguing about who splits which bits, how fast and in what ways.
    And also stop taking about Digital groups.  Who cares what group is
    what. If we started focusing on how we were going to SELL stuff we'd
    all be growing!
    
    What makes customers WANT to buy our stuff (i.e. applications) is
    important.  Speed is important too.  But if you can't do what you need
    to on the box (via packaged software) who cares. (i.e. if Alfer didn't
    run  Windows/NT but only OpenVMS...NICHE MARKET)
    
    
    Aarugh,
    j

    If the mythical customer in this discussion really want an Intel box,
    Digital can and should sell him one, if he wants an Alpha, digital can
    and should sell him one.