Conference 7.286::digital

    Rule #3 of Bill's rules:

    "The market works - keep in close touch; move with it."

    If anyone in our industry is an example of doing that to near
    perfection, Intel & Microsoft are.  There will be a market friendly 
    growth path, when the time comes.

    Husted smoke.  It's always amazing at the propensity of some of these 
    "market" watchers to predict disaster so far ahead, with the absence of
    objective data, toward companies that owe their current industry position 
    to unparalleled mastery of the marketplace.
    
    My $$$ is on Grove and Gates.

    Uh, correction...that's Hastings, not Husted 8~).

> RISC chips like the P7.  That would mean, Hastings said, that 
> Windows 95 could be a short-lived operating system.

This is big news?  Someone from Microsoft (Allchin? I forget.
Could have been anyone except Brad Silverberg, I suppose)
has already said as much in the press.  I think his
prediction was about 3 years.  The long term direction
for Windows is Windows NT.  Windows 95 exists because
too many people can't afford the hardware needed to 
run NT; but hardware gets cheaper. 

Even the P6 is a better fit to NT than to Win95.  That's because
16-bit code slows down the P6, and some significant components
of Win95 (probably USER) are still 16-bit code.  Not so with NT.

  The trade rags have also carried stories describing Intel as
  developing two different "P7"s: One is the obvious follow-on
  to the x86, Pentium, and P6, and uses all the same performance
  through micro-level RISCYness tweaks. The other is the H/P
  VLIW x86+PA/RISC thing.

  Whichever gets to the gate first with the most gets the
  official monicker of "P7".

                                   Atlant

    gosh, two P7 projects.  Sounds like something KO would've done.
    
    Mark

Intel is now marketing the P6 strictly toward SMP NT server.  The Win 3.1 
and WIN95 users are expected to migrate to the P55, running at 150Mhz and 
180Mhz.


							-Paul

    I see where Motorola just announced a new semiconductor facility near
    Richmond, Virginia.  I believe the cost of the facility is going to be
    $3.0 billion and employ up to 5,000.  It seems like the article
    mentioned that PowerPC chips would be built there.  This could be a
    move by Motorola to have the production capacity to offer the desktop
    market an alternative to Intel when the x86 architecture hits the wall.
    

    RE: .3
    >> RISC chips like the P7.  That would mean, Hastings said, that
    >> Windows 95 could be a short-lived operating system.
    
    >This is big news?  Someone from Microsoft (Allchin? I forget.
    >Could have been anyone except Brad Silverberg, I suppose)
    >has already said as much in the press.  I think his
    >prediction was about 3 years. The long term direction
    >for Windows is Windows NT.
    
    
    	Bill Gates said as much on Larry King live. When ask about having
    to purchase Windows 96, 97 as follow-on's to Windows 95, Bill said...
    Buy Windows 95 today and you will be set for about 3 years. 
    
    	3 years from now your WNT choices will be "Windows NT Workstation"
    and "Windows NT Server" or the same WNT choices you have today.

    For those who have not been following the the various P6 articles, some
    further data might add some insight. The current P6 chips are made on
    Intels 5 micron process line and have shown themselves to "test poorly"
    onbenchmarks that involve older 16-bit code. The official Intel line is
    that the design engineers made some trade-offs, based on the
    assumptions early in the P6 design process. Those assumptions had to do
    with the timing and marketplace acceptance of WIn95 and subsequent
    32-bit follow-ons and the attendant movement of the application market
    ot 32-bits. The delays in getting WIN95 out have come back to haunt
    Intel more than Microsoft. The design trade-offs involved the decision
    not to use up transitors for executing some complex 16-bit
    instrucutions. Rather those are emulated on the P6. We all know how
    fast that can be - even the Alpha runs slow when emulating the Intel
    CISC instruction set. 
    
    So, Intel has a temporary problem, until the world moves to 32-bit
    processing and apps, which the P6 does run about twice as fast. They
    aslo have an interim solution - a faster chip implementation, based on
    their .3 micron process line. That solution will come on line early
    next year. The .3 micron chips will be clocked faster, so that even
    though it is still an emulation, it won't be as noticiably slower than
    current Pentiums. In 32-bit mode, the .3 micron P6 will scream and will
    easily keep up with the PowerPC (although not the Alpha). At the same
    time that Intel moves to the .3 micron process they plan to introduce a
    dual cavity model with twice the secondary cache, as well as
    introducing a model without the dual cavity design and no secondary
    cache. 
    
    One of the more interesting charts that Bob Palmer showed in his recent
    Fianancial Analysts briefing was the chart that showed the P6 and P7
    staying right with the PowerPC and MIPS chips and beating the SUN chip,
    but unable to come close to the Alpha chip. It looks like we have the
    better mouse trap. Now if the world only knew the way to our doorstep.
    
    -OFWAMI-  

    
    	Or we have the brains, and guts, to "light" the path to that door.
    
    	But then, when is the last time Digital spent money doing real
    marketing as opposed to no-content advertising and feel good road
    shows?
    
    
    		the Greyhawk

    Have no fear: Caldwell and the Digital Semi pricing committee will
    keep EV5 (the only one of the three Alpha models faster than Pentium
    is *now*) at such a completely rediculous price that few inroads
    into the mainstream will ever be made.
    
    On a related note: there's still time to get FAB6 ready for Spooky
    World; might as well use the excess capacity for something. 

  The trade rags have been carrying similar articles, and this
  really is supported by the earlier thinking in RISC design.
  While Intel has done an enormous job of adding RISCy features
  to a CISC architecture, there's only so much they can do.
  Without the higher-level knowledge that only the compiler
  has about the structure of the program, it takes exponentially
  more work by the chip's micro-architecture for diminishing
  returns.

  P6 offers very good performance on some instructions and med-
  iocre performance on others.

  Now this will be nay-sayed by those who have an extreme vested
  interest in the x86 CISC architecture. But it's still a pretty
  good bet to be true. PowerPC is the obvious heir to the throne
  of "next humongous microprocessor success", but some smart
  marketing might give Alpha a shot at this. Do you suppose we'll
  do it, or go down with Wintel's ship, out of not wanting to bad-
  mouth our PC business?
                                   Atlant

    .12> On a related note: there's still time to get FAB6 ready for Spooky
    .12> World; might as well use the excess capacity for something.
    
    Ummm...  Howzabout "Wear an ALPHAgeneration mask this Halloween...  
                        They're so fast you won't mind how small they are!"
    
    Do I get a reward for helping our marketing effort along?? :-)
    
    

    BusinessWeek, 9/11, p. 50
    
    IBM MAY PULL THE PLUG ON POWERPC
    
    After Two Years Of Bungling, sales of PowerPC
    computers are below IBM's modest hopes.  That's
    why - in the next few weeks, insiders say - IBM
    will decide whether to pull the plug on its Power 
    Personal Systems Division, which makes PowerPC,
    folding its products and development efforts into 
    IBM PC Co. or another unit.

I just came from a meeting in which someone said her group needs support
for a couple of PowerPCs.  She really meant PPC-based Macs.  Then she and
another guy added, almost simultaneously "well, they aren't really
Macintoshes either...."

___Pete

>                      <<< Note 4114.15 by HERON::KAISER >>>
>
>...PowerPCs.  She really meant PPC-based Macs.  

If this is what she meant, then they ARE "real Macs."
Apple has made the transition to PowerPC as the Mac processor.
The next version of the operating system (Mac OS version 8, code name Copland,
due out in mid-1996) is expected to support ONLY PowerPC-based Macs.

There are other PPC-based platforms.  The most contentious right now
is called "CHRP" (Common Hardware Reference Platform), which is a PowerPC
platform that is targeted to eventually run ALL the major desktop operating
systems: Mac, some flavor(s) of Unix, Windows NT, and probably more.
It's possible to have PowerPC "Mac-like" things that aren't Macs.
Could these people have CHRPs?

- tom]

> It's possible to have PowerPC "Mac-like" things that aren't Macs.
> Could these people have CHRPs?

  Or PREPs (the earlier, pre-CHRP PowerPC Reference Architecture)?
  IBM is currently selling PREP-compliant systems running AIX.

                                   Atlant

>I just came from a meeting in which someone said her group needs support
>for a couple of PowerPCs.  She really meant PPC-based Macs.  Then she and
>another guy added, almost simultaneously "well, they aren't really
>Macintoshes either...."

    Perhaps this is like saying that OpenVMS on Alpha is not really
    a Vax.  Except that Apple didn't change the name of their operating
    system when the converted to a new processor chip.  And they
    dont differentiate between the OS and the Hardward...its all a Mac.

    But anyone who does not think a PPC based Mac is 'really' a Mac
    has never used one.

>    system when the converted to a new processor chip.  And they
>    dont differentiate between the OS and the Hardward...

	Hmm, you mean Apple calls the hardware "System 7" as well?

I would venture to say that the majority of the MAC's users don't even KNOW 
what System 7 is! Like I think .18 was trying to say, Apple's marketing 
emphasizes the benefits of the overall product (a valuable productivity 
tool) 
and not its components (OS, hardware, periferals). 

        C'Ya,
        Wayne

    BTW, Its "Mac" or "Macintosh", but not "MAC".  It is not an acronymn.
    
    On the contrary, I would think that most Mac users are aware of System
    7 and beyond.  After all, they are selling System 7.5 for $99 to
    the existing Mac community.  I think that most Mac users know what
    "System 7" is since a large part of the software available for the Mac
    now requires System 7 to function (or have limited functionality on
    System 6).
    
    				-John

How many times have you seen/read/heard Apple use the phrase "System 7" in 
ANY advertisment? 

For example, look at Apples new TV ads attacking the ease of use claims of 
Windows 95. I don't think they mention the name of the OS at all. Again, 
the benefits of the complete product is emphasized and not its components. 
This has everything to do with the Mac's target audience, most of whom 
couldn't care less about what's "under the hood".

	C'Ya,
	Wayne

The name of the Macintosh operating system is now MacOS.  System 7 then refers
to a specific version(s) of that operating system.

Paul

    re: .22 -
    Amen- Apple is fighting to retain their brand image as STILL the easiest
    system to use and install.  And that is the company promise - not the
    operating system. It may do no better for them than hold market share,
    but they still have the best out of the box/ease of use consumer 
    story to tell, although the lead is eroding.

    re:-1
    
    ...and so is thier comittment to the OS as they try selling it or
    themselves to Big Blue.
    
    ..Larry