Conference 7.286::digital

What would we do to "embrace" Linux that we're not doing today?
Can we perform those activities without destroying its low-price
advantage?

	I was under the impression that we encouraged the port
	to Alpha, offered technical assistance, and even donated
	hardware to the cause. Our support at an engineering
	level has been quite good I suspect.

	Do you want us to advertise too? That seems to be already
	being done in that marketplace. If you want us to start
	marketing Linux, well, can we start with marketing our
	own stuff first?



								mike

    RE: -.1
    
    The Alpha (native) porting campaign was big a few years
    ago, but it has since peetered out, to the best of my
    knowledge. This is unfortunate.
    
    Although, I hear that yet another campaign is now
    in the works...
    
    George

	With all the investments we have made in several
operating systems and software to run thereupon, what is
the advantage of doing it once more?  It is my impression
that this market is primarily interested in price - meaning
free software for everything, and bargain basement hardware
prices.  And no support contracts.  Where should we plan to
make money in that arena?

	Do you believe that we can shave our profit margins
on Alpha Linux systems such that:

	1. We have an attractive price for that market
and
	2. We can still make enough money to break the current
death spiral?

	If you are going after volume, how about dropping the
price on any/all of our current software systems to that SAME
price point, and just offer what Linux would offer - no support?
What is to prevent us from doing this - well, probably the fact
that we know that our volume will not be NEARLY enough to
generate the revenue needed to survive with the lowered price.

	My point is that if you are going after a volume market
which is primarily interested in low price, it is perhaps
better to drop our current prices and offer no support than to
invest in yet another system and then offer no support.  Or do
people really believe that Linux is fundamentally more attractive
as a software system than Digital Unix (or our other offerings)?
If so, what are the reasons?  Should we be giving away our sources,
too?

	Remember, if you sell directly to the Linux market, you
have to believe that you are fundamentally just a hardware vendor.
And you have to plan on obtaining nearly ALL your profits from the
sale of hardware.  I think that's wishful thinking.

Cheers!
Dave Eklund

    re: .1, .2
    
    We could start with some simple things:
    
    	a) Press Release(s) stating that we are 100% behind the Linux/Alpha
    	   effort and we will begin listing preconfigured Linux/Alpha
    	   boxes, which can be ordered through a supplied list of
    	   Linux/Alpha distributors.
    
    	b) Become willing to stand up publicly and announce that Linux is
    	   our 4th O/S.  This is not the same as the admirable work that
    	   a handful of our people have done to promote the Linux/Alpha
    	   development.  This is the Marketing stuff of letting people know
    	   that the Corporation, not just engineering, identifies Linux as
    	   an important part of our O/S offerings.
    
    	   Linux/Alpha units at trade shows next to our other offerings.
    	   VPs speaking in public about the price/performance of
    	   Linux/Alpha.  Press Releases that mention Linux whenever we
    	   mention the other 3 O/S offerings.
    
    	c) Partner with Linux/Alpha OEMs (see [a]) to have them supply
    	   post-installation support.  Many of them already offer support
    	   contracts to customers; this can continue under the umbrella
    	   of a Digital blessing.
    
    	d) Dare to make the name Digital synonymous with Linux.  There are
    	   a HOST of corporate people out there who are actively
    	   considering Linux solutions.  If Digital is willing to stand up
    	   in the marketplace and proclaim "We do Linux and we're proud of
    	   it", we run the risk of actually finding corporations who are
    	   willing to jump on board and deploy Linux solutions now that a
    	   major player in the industry is standing behind Linux.
    
    	e) Actually give sales reps INFORMATION, including part numbers,
    	   prices, and distributor contacts for funneling Linux/Alpha
    	   systems.  And, dare I say, allow them to get some form of CREDIT
    	   for completed sales (now THERE'S a rathole waiting to get
    	   started!).
    
    	f) Then there's the dreaded "A" word: Advertise.  AXEL::FOLEY
    	   suggests we should advertise our own products first.  While I
    	   can identify with the sentiment, I must remind you that we
    	   seem to be spending our time and money promoting a certain
    	   Proprietary O/S that is not ours.  Why stop now?  The WNT
    	   game is based on the premise that even a small fish in a
    	   large enough pool can do well (and maybe become a bigger fish).
    	   We have the opportunity to become a large fish in an expanding
    	   pool.  Played correctly, this could yield SERIOUS revenue for
    	   Alpha-based systems and servers.  Plus, it could give us
    	   additional market presence that could seriously enhance our
    	   image as a server & Internet company.
    
    In short, we can play out a market that no other large company is
    really pushing.
    
    We can:
    
    	1) become THE company for low-cost, high-performance servers and 
    	   Internet servers.
    
    	2) increase our presence in the marketplace, aiding our position
    	   to land non-Linux/Alpha sales.
    
    	3) increase our revenue from Alpha sales.
    
    	4) increase our chip volume, thus (hopefully) lowering costs, and
    	   increasing our ability to compete in ALL product sets.
    
    -- Russ

    re: .4
    
    We are already competing in the Linux marketplace with our Alpha boards
    being sold through OEMs/VARs/whatever.
    
    What we need, IMHO, is to give the marketplace a PUSH.  We supposedly
    are making money every time we sell hardware to these vendors who sell
    Linux/Alpha systems (if not, we're in deep sneakers already).  What's
    needed is to increase the demand for these existing products.
    
    Imagine, though, if we could legitimize the Linux marketplace -- it's
    still considered by many in the corporate world as a hobbiest's O/S.
    We have companies looking for low cost, open standards, high
    performance solutions all the time.  The problem is that most of these
    sales go to our numerous competitors.  We could compete at a level
    that could establish ourselves as the preferred low-end vendor as well
    as a preferred high-end vendor (Digital Unix).  We could even offer
    Linux to Digital Unix migrations for growing companies!
    
    Like it or not, our current focus on future profits is based on selling
    Windows NT for Microsoft, supporting Windows NT for Microsoft, selling
    iron for Windows NT for Microsoft.  If that makes sense, then why not
    do the same for Linux?  Linux can allow us to compete at a lower
    price tag (or higher margins, take your pick) while building a name for
    ourselves which isn't overshadowed by the names Intel and Microsoft.
    
    We can improve our financial position AND our presence in the
    marketplace.
    
    -- Russ

    re .5
    Isn't the continuing falling price of Alpha systems being auctioned
    off bundled with Linux (www.onsale.com/category/inv/00041543.htm)
    a pretty good yardstick that there ain't a helluva lot of demand?
    K

    
    Part of Sun's success has been due to the tremendous mind share of the
    graduates from our universities. In fact, much of the mind share for
    UNIX is due to the "what I learnt first" syndrome. When VMS was
    supplanted in Universities many of the graduates no longer had an
    affinity (now there's a word) for Digital or a respect for our systems.
    Universities brought on price. 
    
    Sun was cheaper, had higher performance at the time, and the OS was free. I
    dont believe for one minute that they:
    	a. made much money from the box sales
    	b. believe that it was not a wholly worthwhile activity
    	c. dont think it was responsible for a lot the tremendous follow-on
    		goodwill engendered in business for them
    
    This is not immediate, bottom line stuff, the payback starts when the
    student leaves and starts work. NT is starting to have an impact
    on that base, but Linux is becoming the lingua franca (you can even
    run it on a 386 and have a semi-usable system, 8^). 
    
    I dont believe many of the Sun's sold were high-end systems. We could
    produce a minimal Alpha system without too much pain. The key issue is
    not the OEM boards. We want Digital's name associated with the systems.
    Cheap packaging, minimal memory (32 meg EDO), IDE drives, 1Meg stock
    standard PCI graphics card, stock standard PCI network card, cheap CD
    rom, fast Alpha. Minimal, but totally extensible.
    
    A real low end PC-style config, source it from Samsung if our engineers 
    pride would be hurt by such an non spec system.
    
    Call it the Digital Torvald, and have that on the box. Make the box
    distinctive 8^).
    
    And realise, this is not about margin now, its about:
    	1. market share 
    	2. marketing in the broader sense
    	3. demonstrating committment
    	4. capturing mindshare and loyalty prior to it being totally wiped
    		out by the realities of computing as a business
    	5. long term profit
    	6. upwards sales into parts of unis where they will pay money
    
    
    
    
    - John
    
    

    Don't get me wrong, but...
    
    what would our "partner" (you know, uncle Bill's behemoth) think of us
    offering a hardware/os combination that undercuts by a significant
    amount of dollars the comparable combo with their baby?
    
    H
    
    (a new Linux sys mangler on his home DECpc MTE d2/466 - DEC and Linux - 
    a _great_ combination) 

    re: .7
    
    Isn't the presence of vendors selling commercial Linux/Alpha systems a
    pretty good yardstick that there IS some demand?
    
    The liquidation of 166 Mhz systems, sold with 0 mb of RAM, no hard drive,
    no monitor, no keyboard, and no CD-ROM to load the supplied Linux O/S 
    on CD is hardly the justification for anything.  The boxes described 
    are entirely crippled and will take a significant investment (playing 
    by our rules, whatever they were for these boxes) to even make them 
    bootable!
    
    Your logic could easily be expanded to Alpha systems in general.  If
    these (crippled, old) Alpha systems are selling for a few hundred
    dollars, you'd better get on the horn to the heads of Digital
    Semiconductor and tell them to pull the plug on all this Alpha stuff
    immediately!
    
    No, this doesn't fly.  These boxes were hot stuff when Pentiums were
    running at double-digit speeds.  They're simply no big deal now.
    
    -- Russ

|    Don't get me wrong, but...
|    
|    what would our "partner" (you know, uncle Bill's behemoth) think of us
|    offering a hardware/os combination that undercuts by a significant
|    amount of dollars the comparable combo with their baby?

     Since when did Bill care about our thoughts? This is survival, man.
     We're a hardware company, doing software to leverage our boxes. If
     someone else can do the software, fine. Anything to pump up the
     hardware volume. I doubt the Linux crowd would take away any significant
     business from other Digital groups, at least not for a while.

     >Per

    
    
    If the linux crowd is able, even happy, to run on x86 cpu,
    why would a large share of them be willing to pay more to use
    an alpha cpu?  Seems clear they buy on price not value - having
    opted for a free unsupported O/S. 
    
    Are we to believe that individuals who buy a free O/S that runs fine
    on x86 would pay more for an alpha based system -- when evidence is 
    all around us that companies with deep pockets and more demanding 
    computing requirements will not pay more for alpha? 
    
    
      
    
    
    
    

    re: .12
    
>    If the linux crowd is able, even happy, to run on x86 cpu,
>    why would a large share of them be willing to pay more to use
>    an alpha cpu?  Seems clear they buy on price not value - having
>    opted for a free unsupported O/S. 

    Yes and no.  Who says that Linux represents less of a value?  I've been
    using it for years and my impression is that it is MUCH more reliable
    and valuable than any Microsoft operating system I've used.  Not to 
    mention the fact that it is multiuser, supports X windows and TCP/IP, 
    allows easy remote management of the box without add-on software, 
    subscribes to POSIX standards, doesn't hang or GPF every couple of
    hours, etc., etc. ...
    
    And, what about this "unsupported" bit I keep hearing?  I can purchase
    Linux support from any number of sources.  Yes, it doesn't include the
    fine level of Microsoft support so many people have become accustomed
    to (e.g., "Yes, you have a problem.  No, we won't fix it.  Wait until
    the next upgrade.  Maybe that will fix it." *click*).  Instead, you may
    have to pay one of the Linux consulting firms to fix the source code
    for you.  Or, if you're not in a hurry, talk about the problem on
    Usenet and wait a few weeks until someone decides to fix it for free.
    
>    Are we to believe that individuals who buy a free O/S that runs fine
>    on x86 would pay more for an alpha based system -- when evidence is 
>    all around us that companies with deep pockets and more demanding 
>    computing requirements will not pay more for alpha? 
    
    The question comes back to the issue of value.  Companies will pay
    money for our systems if we are perceived to offer greater value.
    We are at a point in history where many companies are purchasing
    Internet servers to host their own Web sites.  They see the need to get
    a sufficient Web presence, but they don't want to spend gobs of cash on
    a bleeding edge system.  Alpha/Linux could be an excellent and natural
    fit here: tiny (or zero) cost for the necessary software and a
    reasonable cost for a screaming piece of hardware that can handle real
    volumes of Internet traffic.
    
    There is something to be said for software that runs on both cheap x86s
    and higher-end machines.  There had better be or we're doomed with our
    WNT strategy!.  The problem is the Windows NT situation is DEFINED by
    the Intel architecture.  We are constantly fighting uphill because
    we're "not x86 compatible".  FX!32 should help things, but there still
    is a FUD (fear, uncertainty, and doubt) problem that we have to resolve
    with potential customers.
    
    The Linux world is still young and not so tightly restricted to the x86
    mindset.  Yes, it runs on x86.  And Alpha.  And Sun.  And Mac.  And
    other platforms, with more in the works.  But there still is no
    champion among them.  Digital has a chance to become that champion.
    
    Consider something else:  the Linux world is based on open standards. 
    The Microsoft world is based on Microsoft standards.  Linux customers are
    less likely to demand Intel compatibility, since the Linux world is
    designed to be multi-platform.  Like it or not, the Microsoft world is
    Intel-centric, because it has been historically based SOLELY on the
    Intel architecture.
    
    So, in the Linux world, we can compete on REAL VALUE (i.e., the ability
    to do more work faster for less money) rather than on PERCEIVED VALUE
    (i.e., "yes, you're fast, but you're not really x86 compatible").
    
    Can you imagine a marketplace where Alpha stands as the benchmark
    platform?  A marketplace where people say, "that's a nice box, but is
    it as fast as an Alpha?".  The ONLY growth arena that I can see where 
    we could rise to be the #1 hardware platform (in mindshare, at least)
    is Linux.  The Microsoft hardware platform of choice will continue to
    be Intel for the next several years at least; the dominant mindshare 
    of Intel and its clones will take years to unravel (if it unravels at all).
    
    Note also that I'm not saying we start by throwing zillions of dollars
    at this effort.  We could do a lot just by dispatching a few of our
    hundreds of VPs to various corners of the globe holding press
    conferences and giving magazine interviews about our new official
    corporate support for Alpha/Linux.  Trade rags are already starting to
    give occasional press to Linux.  The interest in this operating system
    is growing daily.  If we can just deliver some warm fuzzies to the 
    corporate world, we might just find ourselves riding a wave of free 
    press coverage that might well equal the value of whatever investments 
    we make initially.
    
    The Linux world has already begun to gravitate to us based on the value
    of Alpha.  A wise company would encourage the market to grow, while
    fostering a sense that we are a, if not THE, defining platform for this
    marketplace.
    
    -- Russ

    Never one to have the same point of view but...
    we should drop Unix and adopt Linux since they cannot coreside in
    marketing terms
    we should probably package our own version of Linux
    we should definitely provide Linux support services and hot line
    we should probably consider selling enterprise Intranet servers running
    specific server functions(like mail hub or certificate servers) but
    based on Linux OR NT
    we should probably base our NC on Linux/Java (if we don't already).
    Added value of all this?Services and support
    Dear Unix Buffs - do not kill the messenger please!
    ;-)

    Since a very good friend of mine is employed full-time by DIGITAL
    porting LINUX to Alpha systems and is a contributor to the Linux
    journal, I know you'll find that there is a lot or work going on with
    Linux in the corporation.
    
    /a

	No, the Alpha Linux work has not petered out.  See 
	http://linux.reo.dec.com internally and 
	http://www.azstarnet.com/~axplinux for web sites.  It
	is more than mainstream, we're quietly selling reasonable
	numbers of Alpha Linux boards...

	Dave (aforementioned friend of Andy's)

>    Part of Sun's success has been due to the tremendous mind share of the
>    graduates from our universities. In fact, much of the mind share for
>    UNIX is due to the "what I learnt first" syndrome. When VMS was

I've been out of school for a couple years now, but while I
was there, Linux was the OS of choice.  These people graduated
and went on to suggest Linux as the OS of choice at work.  They've
had varying degrees of success at swaying management, but as
I've mentioned before, several of them now use Linux at work.

As for price/performance... yes, these projects usually want the
best price/performance ratio.

John

|    we should drop Unix and adopt Linux since they cannot coreside in
|    marketing terms

	I take it you mean in "Digital" marketing terms. Did Microsoft
	drop Windows when they had NT? Are they dropping Exchange now
	that they are moving to something "better" for MSN?

	We don't have to market Linux as a replacement to UNIX. Let Linux
	find its own way, but continue to be very supportive of the efforts.
	Consider offering seed units of new low-end Alphas to companies
	like Red Hat. Encourage our Alpha OEMs to do the same. The moment
	we take over Linux 
	
|    we should probably package our own version of Linux

	I think we would be much better off partnering with someone else.
	Re-sell someone else's, like Red Hat. Allow customers to buy Alpha's
	preconfigured with Red Hat's distribution. The reason is simple.
	We are a hardware company. With Red Hat's success comes ours.

	>Per

I can believe that Linux has caught on in colleges.  A son of my parents'
neighbors just adores it, and I think he's got a pretty darn good resum�: 

	http://www.clarkson.edu/~otericj/linux/resume.html
				/AHM

     why linux !?       -----> look at Postscript!!
    
    You can buy a "Multia" (233Mhz/EV4  Zlxe-Grapics onboard,
                            networkinterface onboard, scsi onboard
                            1 half-length pci slot on it, 500mb Disk
    			    4-Memoryslots, cache from 512K to 2MB)
    
    this System will cost about sfr 750.-  (nearly $600.-) 
    (Distributor = "B�rgi"(BUERGI) in Switzerland).
    
    This "UDB" runs originally with WNT 3.51/4.0;
    also with "RED HAT LINUX".
    
    But the clue is: if you do a Firmwareupgrade, it works also with
    Digitalunix; it's not supported but it still works.
    (Some Distributors are selling DUX - Licenses for nearly $1500.-)
    
    --oo--
    
    For a good Profit do following 5 steps; 
    Do it yesterday, and do it all at the same time.
    ----> Sell our "Personal (Alpha) Workstations" not only with
          "WNT" and "Linux".
          Sell them also with Digital Unix. 
          Take only $1500 to $2000 for a DUX-License.
    ----> Make a lot of nice an easy Development-Tools around DUX.
          Make a lot of "Trendy Tools" !!!
    ----> spend our different Softwarepartners fully DEC-Support, to move
          their Software on our Operatingsystems. (DUX,WNT,VMS)
          (This will cost our Company only few Slots. This put  a lot of
          knowhow into Digitals Support and Databases !!! --> better Quality)
    ----> Digital must be on all important SW-Application-Platforms.  
          1 out of 3,  prefered  HW & OS Partner. 
          (He Marketing, whats about Acrobat-Reader on DUX? One of our 
           Costumers (110 DUX-Users) will buy a SUN-Server. 
           Costumer means, a WNT-server does not really perform with 100
           Users.) 
    ----> Reduce the differnet Workstation Modells; (1 desktop, 1 Tower)
          (save a lot of SpareParts around the world.)
    
    Best Regards , helmut
    
    ps.: Linux is nice; Help Costumers play with it on ALpha !!! 
         Later, if they like to have "real" support, so they could do a 
         "CHEAP & EASY-Update" to Digital UNIX.

this may be the wrong place to ask, but that's never stopped anyone else 
before 8-)

why all this talk about Linux and no mention of FREEbsd?  Aren't they kind of
sort of the same thing, from a freeware UNIX perspective?  I know Linux is 
certainly more pervasive, at least in terms of books written and apparently in
the numbers of users, but I never understood why.  Does Linux support a broader
set of hardware?

Just curious... 

-mark

    Excellent comments about "seeding the Universities". IMO, our failure to
    "seed" the universities properly is one of the main causes of UNIX
    being so popular at all. 
    
     Had we properly stocked the colleges with cheap/free hardware and FREE 
    software (ala BSD) I think the world would be a much different place. We 
    STILL give bigger discounts to big business, without regard for the desires
    and preferences of the "new guys" that start work every year. I've seen
    it many times.
    
    .mike.

FreeBSD doesn'ty run on Alpha (but NetBSD does).

    re: .21

>why all this talk about Linux and no mention of FREEbsd?  Aren't they kind of
>sort of the same thing, from a freeware UNIX perspective?  I know Linux is 
>certainly more pervasive, at least in terms of books written and apparently in
>the numbers of users, but I never understood why.  Does Linux support a broader
>set of hardware?

    A tongue-in-cheek response, attributed to Linus Torvalds, the man who
    gave birth to Linux:

    > > Other than the fact Linux has a cool name, could someone explain why I
    > > should use Linux over BSD?
    >
    > No.  That's it.  The cool name, that is.  We worked very hard on
    > creating a name that would appeal to the majority of people, and it
    > certainly paid off: thousands of people are using linux just to be able
    > to say "OS/2? Hah.  I've got Linux.  What a cool name".  386BSD made the
    > mistake of putting a lot of numbers and weird abbreviations into the
    > name, and is scaring away a lot of people just because it sounds too
    > technical.
    (Linus Torvalds' follow-up to a question about Linux)

    
    The _REAL_ answer might be found in the following quote:

    
    "The most important design issue... is the fact that Linux is supposed to 
    be fun..."
    (Linus Torvalds at the First Dutch International Symposium on Linux.)

    
    -- Russ

    re .18
    Per,
        I don't see the way we can differentiate Unix vs Linux.All the
    major things (like SMP ) can be done on both and support of a hardware
    environment(drivers for boards etc) is of dubious marketing value.We
    can't pull the trick of saying NT-is-the-big-brother-to-w95 or a server
    to a client (UNIX the server and Linux the thin client??Makes sense but
    I don't think this is our strategy..).So,for me it comes down to a
    choice - in favour of Linux is the near-zero cost of product
    development and time to market.
         Incidentally,great quote about IPv6 not being ready since there is
    not a Linux version yet..
    	Our own version?You're right that adopting someone elses makes good
    sense but ours would also have a Digital support contract and hot line
    and it would be more serious than a simple resale effort.It's true that
    any version of Linux on a CD is no longer the Linux-spirit since it
    floats in time so a Digital Linux as apposed to a RedHat would be OK.
    
    My own feeling is that we are having a very hard time adjusting to the
    new world where software is free and proprietary is valueless (like
    patents for software design - good grief,are we mad?) and service is
    the only true differentiator.I only wonder how much longer we can
    go one like this.
    

    > > Other than the fact Linux has a cool name, could someone explain why I
    > > should use Linux over BSD?

Linux:  long i or short?

>    > > Other than the fact Linux has a cool name, could someone explain why I
>    > > should use Linux over BSD?
    
>Linux:  long i or short?
    
    The short answer:  usually.  8^}
    
    There are three (!) schools of thought on this.  There's an .au
    file floating about where Linus Torvalds pronounces "Linux".
    To the English-speaking (American?) ear, the first syllable sounds
    closest to "lee", as does the first syllable in his name.
    
    Some English speakers have decided that as the name "Linus" would be
    pronounced "Lie-nus" in English, the O/S should be pronouced
    "Lie-nucks".
    
    Others (in what I perceive to be the currently dominant camp) say that
    "lee-nucks" sounds closer to "lin-nucks", hence the use of the short
    "i" sound.  (This is the official stance of the Linux Documentation
    Project, last I checked).
    
    Others, in frustration, have taken to doing their best Linus Torvalds
    immitation and go with "lee-nucks".
    
    Bottom line:  however you pronounce it, Linux is a blast! (and
    infinitely more useful than any Microsoft O/S I've touched to date!).
    
    -- Russ

I know enough people who say Linux is a fantastic personal-owner operating
system to believe that it may be -- for some people.  But for those of us
who want to ignore the operating system as much as possible and think of our
computers in terms of applications and work accomplished, how is the Linux
environment better than the MS Windows family?

- Where is the (relatively) cheap GUI word processor?  ("Emacs" is not the
  answer to my question.  To your question, maybe, but not mine.)
- Where are the (relatively) cheap, easy-to-use, consistent-interface
  spreadsheets, presentation graphics, personal finance packages, databases,
  etc.?
- With Windoze 95 I can (usually) attach a device, boot, and it's there.
  Does Linux make me edit and rebuild the kernel to add a device?  I don't
  want to do that.

As an operating system by itself, Windows-Whatever is pretty messy, and Linux
may well be far superior.  But as Digital keeps painfully learning, the thing
that matters is the total solution -- and applications are the most important
part of that solution.

I'm not denying a reasonable basis for anyone's enthusiasm about Linux.  But
I haven't been able to see how it meets *my* needs for applications and
add-ons.

Will somebody please point out to me what I'm missing?  I'd love to be wrong
in this case...

-Mark

    re .28:
    
�- Where is the (relatively) cheap GUI word processor?  ("Emacs" is not the
�  answer to my question.  To your question, maybe, but not mine.)
�- Where are the (relatively) cheap, easy-to-use, consistent-interface
�  spreadsheets, presentation graphics, personal finance packages, databases,
�  etc.?
    
    Is free cheap enough? StarOffice 3.1 for Linux should be available soon
    (see www.stardiv.de).
    
    I do have the Win95 version (though I don't use it - I use MS
    Office...). It includes all the usual stuff (word processing,
    spreadsheet, presentation, ...).
    
    No, I'm not a Linux fan - I see absolutely no reason whatsoever to use
    it on my home machine (other than maybe "educational" puropses, i.e.
    learning one flavor of Unix). I did have it on my laptop for a short
    time, but never figured out how to get the GUI running on it.
    

    Of course Linux isn't a panacea. Nothing is.

re: .29

Thanks for trying to be helpful -- but your personal testimony brings up all
my reservations:

- the version you recommend (3.1) isn't even available yet
- the version you have (for Windows) you don't use
- you tried Linux and couldn't get it to work (GUI interface)

I'm afraid something free than I can't use is no better than something
expensive I don't buy.

Again, I'm not trying to badmouth Linux for what it is -- just to point out
that it fits a niche only.  A rather narrow niche, in my opinion.  It's *not*
a replacement for everything known to man.

-Mark

|        I don't see the way we can differentiate Unix vs Linux.All the
|    major things (like SMP ) can be done on both and support of a hardware
|    environment(drivers for boards etc) is of dubious marketing value.We
|    can't pull the trick of saying NT-is-the-big-brother-to-w95 or a server
|    to a client (UNIX the server and Linux the thin client??Makes sense but
|    I don't think this is our strategy..).So,for me it comes down to a
|    choice - in favour of Linux is the near-zero cost of product
|    development and time to market.

	I think we can differentiate between the two. First of all, there is
	no such thing as a free lunch. You can benefit from all these goodies
	in Linux by being prepared to take some risks. These risks may go down
	over time, as the public strives to standardize parts of Linux, much
	like what has happened to the internet itself. Today those risks are
	substantially higher than with Digital UNIX. Linux is no panacea.

	Someone might then ask: how should a Digital sales rep know when
	to propose one over the other? Until we have a serious support
	offering for Linux, the answer is clear -- Digital UNIX. But that
	does not mean we cannot have it in the price book. If a customer
	wants Linux, then let them have it, with the caveat that we will
	have to work out a support contract as we go along (no off the
	shelf deal). If Linux does not support all the customer's SW
	requirements on a proposed hardware platform the answer is also
	simple -- Digital UNIX. 

	Coming back to the free lunch. Linux is cheap. But unless you have
	virtually free labour, like a university, your maintenance costs
	may be higher than you would like and your degree of dependability
	may not be as great as with Digital UNIX.

	>Per

    re: .28
    
    The office apps are out there and more are on the way.  .29 cites one
    major effort that is coming to fruition; there are more in the works.
    
    Don't trust software unless you pay for it?  Check out www.caldera.com.
    Buy native versions of WordPerfect, CorelDraw, and ADABAS D.  You can
    even buy Wabi to run most MS Windows apps on your Linux box (if you
    don't want to wait for Wine, the free Windows emulator that's in
    development on the Internet).
    
    Can't live without that ohh-so-cool (huh?) W95 interface?  There's even
    a W95-like version of fvwm (a really useful multipage X windows
    manager) kicking around.
    
    The apps are coming.  The wave is still building.  The crest is NOWHERE
    in sight.  THIS IS NOT A DRILL!
    
    This is NOT just another O/S.  This is the hingepin for a new push to
    realize the many dreams that Unix, the GNU Project, and others had, but
    couldn't deliver.  Truly open software.  Free or inexpensive software.
    
    Will it change the world?  I don't know.  Will it change the way we
    think about computing?  I'm beginning to see signs of that already.
    
    If we want to try to make a place for *US* in this marketplace, this is
    the wave to ride: we already have a market advantage here.
    
    Of course, we can be satisfied with our current status of Bill's
    trained poodle, yipping and dancing on command in the hopes that a few
    crusts of bread thrown from the master's table will be enough to
    satisfy our ever-shrinking size.
    
    There's nothing wrong with getting fed through our "partnerships". 
    But, shouldn't we be aspiring to _WIN_ a market for once?
    
    -- Russ

    re .31: No reason to thank - I didn't even try to be helpful... ;-)
    
    As I said, I have StarOffice 3.1 on my W95 home machine (where I'm
    tyong this right now). I didn't buy it, it was a freebie bundle with
    the Epson printer I bought about a year ago. The little I've  played
    around with it, it's quite ok - but everyone else is using M$ Office,
    and that's what I have to use at least when working on work-related
    documents at home.
    
    They've been talking about the free Linux version for quite some time -
    obviously, it is not their highest priority.
    
    In any case, Linux is worlds away from the critical mass of the Wintel
    platforms. I have to suffer enough Unix in my work, so I don't want to
    expose myself to mores suffering at home...  ;-)
    

    re .33
    Not wanting to come down hard on Linux as well, but I doubt any
    customer is using it in revenue earning, commercial, high availability
    areas, either as a server or a front end.
    
    Greg

>    Not wanting to come down hard on Linux as well, but I doubt any
>    customer is using it in revenue earning, commercial, high availability
>    areas, either as a server or a front end.

I've seen a number of aricles where revenue earning businesses use Linux.
The one that sticks out in my mind was (is) a hospital which had
decided to use PC's runnning Linux to keep a patient information
database.  The DB server was Linux as were (I beieve) the clients.

I saw this article > 1 yr. ago.  I don't remember the source, so please
don't ask. ;-)


...tom

    I think you would quite surprised...
    

>    Not wanting to come down hard on Linux as well, but I doubt any
>    customer is using it in revenue earning, commercial, high availability
>    areas, either as a server or a front end.

I can understand how you wouldn't expect this.  I would never have
expected that places like Fidelity would ever use (any) unix either.

Two "high end" areas I know of:

	A growing startup (20+ employees) that uses linux for their 
 	web server.

	A major piece of a (very expensive) military synthetic theater 
	of war (this is like a simulation with an interface to the real 
	world).

    re: .35
    
    I was involved in a bidder's meeting for a US government agency a
    couple months back.  The subject: the purchase of HUNDREDS of Linux
    boxes to perform a task that would impact every resident in the US
    within days of power-up!
    
    Get used to it.  There are gobs more apps to come.
    
    No one believed MS DOS could be used for anything in business, either
    (some still don't think so...).  But, yet, Microsoft has made more
    money selling that O/S to businesses than we can ignore.
    
    Also, don't forget the timeline here.  Linux was no more than a CONCEPT
    a mere 5 1/2 years ago.  It's currently pushing its way into corporations
    world-wide as we speak.
    
    Note that with things as they presently sit, it is a matter of WHEN,
    not IF, the number of Linux seats worldwide exceeds the number of seats
    serviced by Digital operating systems.
    
    AltaVista says the word "linux" appears 1,617,929 times in its
    database!
    
    			OVER 1.6 MILLION TIMES!
    
    Do you think these are ALL snot-nosed, teenage hackers who use Linux as
    a way of accessing pornography while mommy and daddy are at work?
    
    LISTEN UP!  We are seeing the advance guard of a movement which could
    sweep over us before we ever decide to pay attention to it!  And, the
    truly sad part is that we're in position to LEAD this market and not just
    feed off the leavings of the big dogs in the pack.  But we need to be
    PROACTIVE for once.  You know -- like back in the days when we used to
    make money instead of losing it!
    
    -- Russ

    Sun is scared to death of Linux.
    
    Not for the high end, but for the low-end intel solaris market where
    price has been the differentiating factor. The level of activity and
    quality improvements at each release is scary. And companies like
    caldera do have serious industry players behind them.
    
    The other danger is that there is NO commercial decision driving this
    activity. So, if a feature isnt there, there is no product prevention 
    process to stop it from happening. Pretty soon solaris may get run over
    the top of at the low end.
    
    The red hat RPM system is, IMHO, a quantum leap over the Digital UNIX
    system. Given the job of recognising the amount of stray hardware at PC
    based UNIX has to do, it does an amazing job.
    
    I dont see it replacing anyones desktops soon, except where UNIX is
    needed i.e developers and universities, which were Suns "breeding
    ground". The problem for Sun is that we are seeing our new graduates
    having done their operating system study on Linux sourcecode, not
    solaris. 
    
    In my opinion, this alone is wholly a good reason to get behind it.
    Regards,
    John
    
    

    >Do you think these are ALL snot-nosed, teenage hackers who use Linux as
    >a way of accessing pornography while mommy and daddy are at work?
    
    No, they're more likely to use Windows 95...
    
    Besides, I got 4,000,000 hits for "Windows" in AltaVista, only 700,000
    for Linux (and 2,000,000 for "Unix").
    
    Hotbot brought 517,871 hits for "Linux", 1,972,894 for Windows... (no,
    I didn't check whether they all actually referred to "Windows", the
    various incarnations of various operating systems by Microsoft). BTW,
    VMS had 237,527 hits.
    

    As a unix consultant I am often looking for software for my customers
    Digital Unix systems. It is becoming common to find the best version as
    a Linux version. And generally it compiles and installs on Digital Unix
    just fine. Usually works a lot better than some of the 'supported' stuff
    we put in out distributions. Most recent example is rdist. 
    
    And for an example of the responsiveness of Linux support vs vendor
    support, the Linux fix for the ping-of-death problem was the first one
    available.
    
    We had better learn to live in a world where lots of software is free,
    because it is happening.
    
    And if you want to see real commercial applications of Linux subscribe
    to the monthly magazine LINUX JOURNAL. They have a major article each
    month on some 'bet your business' reference. 

    Which "AltaVista" did you use?
    
    Word count: mvs:115672; vms:390010; solaris:402723; linux:1618007; 
                unix:2401998; windows:5167913
    
    Yea, I'll bet Sun _is_ a tad worried...
    
    

Please see the following auction sale on internet.  Although I don't have
exact numbers but I think most of these HOT boxes are used for LINUX. 
    
Today, demands for LINUX are up.  A friend of mine owns a manufacturing plant
to produce optical fibers.  Three computers used for process control are
running with LINUX.  His opinion is for small-size business ( < 100
employees) as far as he knows people will choose LINUX over other UNIX.
===============================================================================
Subj:	New!! DEC Universal Desktop Box (VX40A-F2) w/External Two Bay Enclosure

This is a multi-part message in MIME format.

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

[Image] New!! DEC Universal Desktop Box (VX40A-F2) w/External Two Bay
Enclosure

[Bid]  [Previous] [Next]  [Index] [Home]

List Price: $4,950.00
Minimum Bid: $695.00
Bid Increment: $5.00
Quantity Available: 5

Auction # 38538

Auction closes at or after Mon Jan 27, 1997 1:52 pm Pacific Time.
Sales Format: Yankee Auction(TM)

Last Bid occurred at Sat Jan 25, 1997 10:36 pm Pacific Time.

The current high bidders are:

  1.  MV of Princeton, NJ, Sat Jan 25, 5:10 pm ($795.00, 1)
  2.  MM of Columbia, MO, Sat Jan 25, 10:36 pm ($750.00, 1)
  3.  FM of Mokena, IL, Fri Jan 24, 10:04 pm ($705.00, 1)
  4.  BD of Billerica, MA, Sat Jan 25, 1:04 pm ($705.00, 1) : "Don't go high
     than this ppl: HD is used!"
  5.  MP of Kamuela, HI, Sat Jan 25, 1:28 pm ($705.00, 1) : "got one
     already; great!"

New!! DEC Universal Desktop Box (VX40A-F2) w/External Two Bay Enclosure with
           32MB RAM, 2GB Hard Drive & CDROM Drive and DEC Keyboard

Price Performance Breakthrough on THE WORKS!

THIS UNIT MAKES A KILLER SERVER OR WORKSTATION UNDER WINDOWS 4.0
WORKSTATION, LINUX OR DEC UNIX!!

The Universal Desktop Box (VX40A-F2) is truly a desktop wonder. Its compact
design integrates the powerful Alpha CPU, SCSI, network, sound, and graphics
subsystems in a single-board package. This is an awesome Windows NT and
Linux workstation! Includes Red Hat 4.0 and a coupon for the full OEM
version of Windows NT 4.0 Workstation for $99. All this unit needs to
operate is a monitor. (check your closet)

Package includes:

   * 166 MHZ Alpha CPU
   * 32MB ram (DEC True Parity - Matched Pairs for Interleaving)
   * External Two Bay Enclosure with one 80 watt Power Supplies and All
     Cables - Holds 3.5" and 5.25"
   * 2GB IBM HD 7200RPM 8MS access time (Recertified -1 Year warranty)
   * Toshiba 2X CDROM
   * 1.44MB floppy
   * 2MB video
   * Keytronics Keyboard
   * Red Hat 4.0 CDROM
   * Coupon for full OEM version of Microsoft Windows NT 4.0 for $99
   * Built-in ethernet
   * CD quality sound
   * DEC 3 Button mouse

              EVERYTHING IS BRAND NEW UNLESS STATED OTHERWISE.

This is a fast workstation for Windows NT and Linux. The same machine was
recently sold by DEC at $1599. It originally sold for $4500.

This machine makes a fantastic:

   * Web server
   * Desktop client
   * 3D rendering machine
   * PC server
   * Mac server

It runs Windows NT, Linux and DEC Unix. It also should run most DOS &
Windows programs with Windows NT 4.0 AND FX32!. It is a true 64 bit system!

Includes a coupon from Starship Computer Company that allows purchase of one
full OEM copy of Windows NT 4.0 Workstation for $99.

SPECIFICATIONS:

* CPU: 166 Mhz DEC Alpha 21066A with 16K internal cache - 166 MHz clock rate
* Secondary Cache: 256KB write-back
* Ram: 32MB. Expandable to 256MB.
* (4) 72-Pin Simm slots (4,16, 32 or 64MB true parity)
* TGA Video w/2 MB VRAM Graphics accelerator with VGA connector
* Resolutions possible: 640X480,1024X768 ,1152X900, 1280X1024 @ 60-75 Hz
refresh rate
* PCI Fast SCSI2 internal/external interface (NCR 810)
* Built in autosensing PCI ethernet (Thick, Thin and twisted pair)
* PCI Expansion - Internal short form PCI card slot
* Duplex 10BaseT
* 1.44MB 3.5" floppy disk
* 1 Serial port with full Modem support
* 1 bidirectional parallel port
* 2 PS/2 keyboard/mouse connectors
* PS/2 2 Button Mouse
* CD audio quality input/output with built-in speaker (Microsoft sound
system compatible)
* CDROM DRIVE!
* Universal Power Supply: 100 to 120, 220 to 240 VAC nominal. 86 to 284 VAC
autorange.

Physical characteristics:

   * Height 7.1 cm (2.8 in) Width 31.7 cm (12.5 in) Depth 31.7 cm (12.5 in)
   * Weight 6.1 kg (13.5 lb).

                     Reminder: Monitor is not included.

Warranty

This item comes NEW with a 1 Year warranty. This warranty applies only to
units that fail as a result of manufacturing defects. Starship Computer
Company is not responsible for any incidental damages that result from
improper installation or operation or damage to other connected components
such as ram, floppy drive, keyboard, expansion card, etc. Starship offers
technical support for the product.

There is a 15% restocking fee within the first thirty days. After the thirty
day period, all sales are final. Windows NT 4.0 Software is not returnable.

Shipping and Handling

Domestic (Continental US) Shipping and Handling Charge: $69.00 per unit
(Alaska and Hawaii may be higher.)
International (Canada and Mexico) Shipping and Handling Charge: $276.00 per
unit

In addition to shipping within US, this product can be shipped to Canada and
Mexico. The shipping and handling charges in those cases are usually about
four times the US rates. No other destinations are supported at this time.

Offered By

Starship Computer Guys
35 Mitchell Rd
Lee, NH 03824
FAX: 415-428-0722

Your credit card will be charged by ONSALE

Sales tax will be charged to CA state residents.

ONSALE accepts only Visa, Mastercard and American Express.

For Order status or other customer service, call Frank Reynolds at (603)
659-2912.

Sales Policies:

Prices reflect a 2% cash discount. All orders will be shipped within 10-14
days from the close of the auction.

Shipping and handing charges will be added to all items sold through,
ONSALE. Shipping charges are for standard delivery in the continental United
States. Please contact for prices on alternative shipping arrangements or
other locations.

Warranty Information:

All of Starship Computer Company's products have been individually tested
and should arrive in perfect shape. If you have any problems with your
products please contact Frank Reynolds at 603.659.2912.All Items that are
offered through the ONSALE service carry a 30-day replacement warranty.
Starship offers technical support for the product.

Starship sells DEC Alphas to customers around the world in the education,
business, entertainment and scientific communities. We have our own line of
Alpha computers and we are a reseller for Aspen Systems and DEC. We also
manufacture the DEC Alpha Farms and the Intel Farms.

Starship was named primarily for what we hope people will do with our
computers, and only partially for Jim, Bones, Scotty & the crew. Starship
Computer Company is a division of American Business Service Corp. , which
has been serving the computer hardware and environmental protection needs of
industry since 1985.

ABSC helped publish the Earth Day 1990 Calendar as well as organize national
environmental photo contests and exhibits. We were part of the team that
developed an environmentally safe method of cleaning circuit boards that is
now marketed internationally.

            6698 - Page Last Updated: Sat 25-Jan-97 22:35:28 PST

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

              Copyright )1997 ONSALE, Inc. All Rights Reserved.        

---------------------------------31074267496389--


% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: from mail13.digital.com by us1rmc.bb.dec.com (5.65/rmc-22feb94) id AA01149; Sun, 26 Jan 97 10:44:13 -0500
% Received: from risc.tre.state.nj.us by mail13.digital.com (8.7.5/UNX 1.5/1.0/WV) id KAA29347; Sun, 26 Jan 1997 10:36:35 -0500 (EST)
% Received: by risc.tre.state.nj.us; id AA14619; Sun, 26 Jan 1997 10:27:22 -0500
% Message-Id: <[email protected]>
% Date: Sun, 26 Jan 97 10:27:23 -0500
% Sender: [email protected]
% From: NJCFS <[email protected]>
% X-Mailer: Mozilla 1.12I (X11; I; OSF1 V4.0 alpha)
% Mime-Version: 1.0
% To: nyoss1::jaung
% Subject: New!! DEC Universal Desktop Box (VX40A-F2) w/External Two Bay Enclosure
% X-Url: http://www.onsale.com/category/inv/00038538.htm
% Content-Type: multipart/mixed; boundary="-------------------------------31074267496389"

re.35:

>    Not wanting to come down hard on Linux as well, but I doubt any
>    customer is using it in revenue earning, commercial, high availability
>    areas, either as a server or a front end.

There's a well known ISP right in your backyard that uses it! 18 months ago 
DECUS in the UK ran a LINUX meeting with a session on commercial use of 
Linux. I'm sure it's moved on since then.

Dave.

    The so-called HotBoxes at www.onsale.com are ex-Multias. They're not so
    hot, but are good enough to play with. For many (most?) people, an
    equivalently-priced Intel box would be a better bargain.
    
    The binary compatibility between DIGITAL UNIX and Linux has already
    been mentioned. I believe it works both ways for statically linked
    executables, i.e. take Linux app and run on (appropriate version of)
    DIGITAL UNIX, take DIGITAL UNIX app and run on Linux (subject to
    licence).
    
    There are/were versions around that were FIPS and POSIX *certified*
    (not just compliant) and therefore eligible for purchase by the US
    Govt. 
    
    My favourite UK retail software catalogue (Software Warehouse) has just
    started featuring Caldera Open Linux in its volume catalogue.
    
    So, is there lots of interest ? Clearly.
    
    Can DIGITAL as she is today (where Palmer's comfort zone appears to be
    "enterprise servers" - which don't sell much because customers are all
    downsizing too) make money from Linux ? I don't really see how, except
    maybe the DIGITAL Semiconductor folks selling chips (boards, etc) to
    Enorex and the like. Which is fine. *Maybe* the next level up the
    market will be happy to pay Digital prices for the Personal Workstation
    family, which would be good. But for the rest of the company, have we
    not enough "core competencies" to focus on already ?
    
    regards
    john

	The Alpha Linux work has been done by/sponsored by Digital Semiconductor.
	We have good links into the major players (Linus, Red Hat, Caldera) and
	so on.  All (yes let me repeat that), all of the Digital Semiconductor
	boards so far produced support/run Linux.  All future boards will run
	Linux.  

	The Linux world is heavily based in education and research.  Linux is
	extremely popular with students and with lecturers.  There are a lot
	of Universities in England that teach computer science using Linux to
	hold together their networks and as teaching aids.  "When you can
	write a Linux device driver, you have graduated".

	The Linux Market is concerned with value for money, what market is not.
	When I buy a car, I buy within a budget and base my buy on consumer
	magazine reports.  Buying a computer is no different.  Strangely 
	enough the buying metrics for buying Alpha WNT versus Intel Windows NT
	is exactly the same as buying Alpha Linux versus Intel Linux.  The
	latest generation of DS alpha boards, the PC164 (ev56 @ 433 MHZ) is
	reasonably priced and selling in volume through Annerex (sp?), Aspen
	and others.  Multias and XLs were ridiculously over priced and 
	undersold (as well as being memory bandwidth limited).  They got Linux
	ported to them (by us) in order to sell them off cheap before they 
	got crushed.

	I am a pragmatist.  I see Alpha Linux as an incremental sale of hardware.
	Looking at the market, I think that we can take the top 5-10% of the
	Linux market.  Our opposition is (surprise, surprise) Pentium Pro, MMX
	and so on.  I never want to see a Digital Linux distribution.  We just
	do not know how to do it, leave that to Red Hat, Caldera, Craftworks
	and so on.  We can sell hardware, we could even sell services.

	I don't want to hear any form of the "how will it affect sales of X?".
	It affects no sales of anything else except positively.  Some folks
	run Windows NT and Linux on their Alpha.  Sometimes, sales of Alpha
	Linux boards reference sales of Digital Unix.  So far as I know we
	neither care about nor affect sales of VMS.

	Lastly and not leastly, I and my other Alpha Linux friends in Digital
	(Jim, Jay, Maddog and others) are regularly asked to places to talk
	about Linux and Linux on Alpha.  They even recently appeared on PC TV
	(great advertising).  We've written articles in the Linux Journal and
	in the new UK magazine Linux World.  We always get a good welcome.
	The same folks who think that Linux is cool (including Linus) also
	think that Alpha is cool.  

	Right back to my job (which despite Andy's note is not totally Alpha
	Linux, I play with StrongARM (another great little processor) too).

	Dave

	ps Go read the web pages, go install Linux and above all go and have
	fun.  This is still an engineering company.

	pps. Gee I hope that I didn't make any grammer or spelling mistakes
	I don't want the thought police to come knocking.

    One thought on the quality of Linux, and associated packages.
    
    Commercial software is produced by people working at companies, subject
    to all the nonsense that Dilbert so well points out each day in the
    newspapers. The programmers work is usually anonymous, maybe your
    initials appear somewhere in a comment line, but otherwise your work is 
    totally unacknowledged.
    
    Linux packages are written by people "for the love of it". The authors
    name is all over it, since packages almost always include source code.
    And the Linux market knows a package by the authors name. Linux
    packages have strong pride of the author connected to them.
    
    So why would you expect 'commercial' software to be better quality than
    Linux packages? 

RE: .35

	One of my clients is using Linux (on Intel) as the Web
	server front-ending their ordering and production tracking
	system.  This box is also a secondary DNS server.  They 
	are looking at adding additional Linux boxes (either Intel 
	or Alpha) in other applications.  The current system has 
	been running for about a year with no problems and the 
	only downtime has been due to one power outage.  Linux is
	out there and it's being used in commercial applications.

	Have a GREAT day!
	Pete Stoddard

    As part of my real life, I am involved in the running and maintenance
    of two TLDs (top-level domains). These are critical to a large number
    of domain name holders. The machine? A 486 DX4-100 running LINUX, and
    it's *FAST*. Anyone who dismisses LINUX as a non-serious OS is a fool.
     
    FWIW, Laurie.

RE: .48 -< Your name in lights >-
    
Before I start, let me say that I have never run Linux, nor (knowingly)
even seen a system running Linux.  So my comments are purely theoretical,
and do not reflect on the quality of any Linux package or developer.

>    So why would you expect 'commercial' software to be better quality than
>    Linux packages? 

One word: testing.

I have done software development for the sheer fun of it, as a hired gun 
to make my mortgage, and as an engineer with Digital.  I have seen many
engineers in my 21 years in this business.  And the good engineers were
capable of producing top-quality, production-ready, bet-your-business
code in any language on any system.

But the good engineers also had bugs in their code.  And since the engineer
wrote it, it was exceptionally difficult for them to find the bug, because
they couldn't see it, and they wouldn't think the way an end-user would
think, or use the product the way an end-user would use the product, because
they wrote the code.

For that you need external testing people.  Call them CSSE, call them 
TQM, call them beta testers, call them anything you want, but you need a
set of people who are independent of the engineer who designed and wrote
the code to run the program through its paces, and expose what was never
even considered or thought of during the design and implementation process.

Microsoft does this brilliantly.  The beta test of Windows 95 was the biggest,
longest, most public field test in history, involved the most people, and
the results prove my point: Windows 95 has had the most features with the
fewest bugs, and by far the fewest serious bugs, of any V1.0 operating system
ever released.  Does it still have some bugs?  Of course, but compare the
level of features and bugs of it with any other V1.0 operating system from
any other company throughout the history of computing.  95 wins hands down.

With commercial code from a recognized vendor you have some assurance of 
the level of testing it has gone through.  With Linux you have none of
that.  Now, the community as a whole may perform this service, but you
have no assurance that any given release of code has survived any level
of testing.  This concerns me.

I agree that engineers who do it for the love of it, and aren't hampered
by Product Prevention Committees (PPC), can turn around bug-fixes much 
faster than commercial companies like Sun or Digital.  But what level of
testing has the bug-fix survived?

Now, because the engineers name is all over it, we will see a Darwinian
process occur, where the user community will say "xxx writes good code,
and yyy writes garbage", and you will end up with a high quality set of
products.  But again, for any given release of a product, you have little
to no assurance of its quality.

Are people using it?  Of course.  But these are Gartners "bleeding edge"
and some "early adopters" companies.  This is a tiny niche market.

Now, should Digital support this effort?  To the level we are now, I think
we should.  If it sells boards, it sounds good to me.  And I agree that it
will not cut into our Windows-NT or OpenVMS sales, because these people
are politically opposed to Linux.  Might it cut into our Digital UNIX 
sales?  Maybe over time, but Digital UNIX is also focused on the high
end, and so I see that as minor if it is there at all.  Should there be
a Digital Linux?  No, that would negate the whole point of Linux.

IMHO, and YMMV, and all that...  But I think that testing as a significant
factor in the quality of software is something that is being missed here.

-- Ken Moreau

    For what it's worth, my wife has been asked to install
    Linux on a PC for the start-up where she works. She has
    some UNIX experience from about 10 years ago. Her general
    perception is that installing Linux and X is incredibly
    complicated. One comment is "the way Linux is organized
    reminds me of the way a 15-year old boy's room would be 
    organized". Documentation is poor to nonexistent, and
    simple things like loading executables from a CD require
    the CD-ROM to be mounted, then dismounted before the
    CD can be removed. 
    
    Maybe LINUX is a great tool for UNIX hackers to play with,
    but a lot of cleaning up needs to be done before it's ready
    for the general public.
    

re: .52
>    simple things like loading executables from a CD require
>    the CD-ROM to be mounted, then dismounted before the
>    CD can be removed. 

I bet your spouse would not think much of Digital UNIX, despite
the mountains of documentation.

Every OS I use actually requires that the CD be mounted to
load files/software/etc, Windows NT does it for me & Digital UNIX
does not (gee, imagine that!).  Making it so that one cannot remove
a CD until it is dismounted is not unreasonable, some software
programs lock the drive until the disk is no longer in use (it is
seen as a security feature, one you'd like if anyone could walk
by the system while you were using it and mess you up by removing
a disk prematurely).  You've chosen a poor example to show the
difficulty of using LINUX.  I'd recommend staying with the 'poor
documentation' complaint, that works for many unix operating systems.

For people with any unix administration experience that isn't a
distant memory, LINUX is not awful.

>For people with any unix administration experience that isn't a
>distant memory, LINUX is not awful.

    OK, I acknowledge that the CD-ROM mount/dismount thing is
    something that Windows does in the background. Here's
    another example: to install X (the documentation makes
    explicit reference to the fact that it's not correct
    to call it "X Windows"!), you need to know a bunch of
    parameters for your graphics chipset. The most common
    way to find these parameters is to cruise the USENET and
    see if anyone who's got your PC's graphics chipset has
    been kind enough to post their parameter settings. The
    documentation warns that incorrect parameter settings
    may cause permanent damage to hardware! Gee, sounds like
    the kind of risk I'm willing to take; what the heck,
    at least I'll learn something about how graphics chips
    work!
    
    Maybe there's an easy way to find these parameters, or
    maybe you can get them from the manufacturer, but the
    fact that you have to do this at all is indicative of
    the problems Linux has. Most people want to use their 
    machines to accomplish things, not learn how they're put 
    together at an atomic level.
    
    - Dave (who's gotten an earful about Linux every night 
      for the past week or so)

> IMHO, and YMMV, and all that...  But I think that testing as a significant
> factor in the quality of software is something that is being missed here.

Personally, I'd prefer that commercial software is developed with
something like Humphrey's PSP method instead of relying heavily
on testing.

Anyhow, there are "released" versions of the Linux kernel.  These have 
been tested by lots of users, developers, etc...  Other versions of 
the kernel are beta tests.

I haven't heard of anyone testing apps or patches to ensure they 
work with specific kernel versions.  But I wouldn't be suprised if
companies that distribute Linux do some form of intergration testing.

    re: .55
    
    Most Linux folks agree that the XFree86 setup stuff is the most difficult
    aspect of launching a Linux box.
    
    If you can afford about $100, you can get one of the commercial X
    servers that (reportedly) solve the configuration problems for you,
    without the hassles of the free server.
    
    You can expect to see further developments in this area in the future.
    
    BTW, which Linux distribution did your spouse use?  Red Hat, Debian,
    Slackware?  Red Hat V3 installation was remarkably painless for me --
    and I installed everything under the sun!
    
    re: testing
    
    Microsoft launches far-and-wide beta testing and is praised.  Linux has
    far-and-wide beta testing and is doubted.  Sounds more like a marketing
    issue than a technical one.  
    
    My Linux desktop stays up for days at a time.  It comes down when _I_
    choose to shut it down, not when _IT_ chooses to shutdown.  I wish I
    could say that about my Digital-issued Win95 laptop!
    
    Again, you want true commercial apps to satisfy your need to belief
    that commercial stuff is higher quality?  Go ahead.  Buy Caldera
    desktop with WordPerfect, etc.  The choice is yours.  With Linux, you
    can choose commercial OR freeware; whatever you're comfortable with...
    You can still save a bundle of cash without paying for the operating
    system, the C/C++ compiler, etc.
    
    re: documentation
    
    Most CDs have gobs of docs on them, although not necessarily in the
    same form as many commercial offerings.  Absolutely need a hardcopy
    book?  Buy one.  You can find really helpful books for as little as $30
    or so.  Personally, I've found the HOWTOs and FAQs more than
    satisfactory for setting up a Linux box.  I broke down and bought
    "Running Linux" by Matt Welsh at a local computer show after I
    committed to Linux as my primary work system.  I figured it was time to
    have a title on the bookshelf to prove I was serious.  8^)
    
    -- Russ

RE: .57
    
>    re: testing
>    
>    Microsoft launches far-and-wide beta testing and is praised.  Linux has
>    far-and-wide beta testing and is doubted.  Sounds more like a marketing
>    issue than a technical one.  
 
There is a difference: with Microsoft you have a formal, incorporated,
financially liable entity which stands behind the software, and which has
demonstrated a strong commitment to find and fix bugs.  With Linux you
have an ad-hoc, informal, nonexistent from a financial standpoint, and 
completely unaccountable group of constantly changing people who have a
wildly varying commitment to find and fix bugs.
    
>    Again, you want true commercial apps to satisfy your need to belief
>    that commercial stuff is higher quality?  

Actually I don't believe that commercial stuff is higher quality.  I believe
that a "true believer" who is committed to the idea of developing software
and sharing it with the world can produce software that is quicker to market,
higher quality, and more responsive to real world users needs than can a
behemoth corporation with all of its overhead.  And I believe that Linux
has many of these people, dedicated to making Linux a success.

The difference is accountability, which was a hard thing for me as an engineer
to understand.  I heard an IS Director one time say that while he believed
that the other company's products could do the job, he was going to buy from
IBM, because he knew that if he called the IBM CEO with his problem, he would
get it fixed quickly.  He had no confidence that this was true with the small
company, and I think this is equally true of Linux.

I guess my only problem with the excitement that you and others have around
Linux is the idea that it is the obvious choice for every computing need.
That it can and should replace every other O/S out there.  That it can
totally satisfy every single customer.

Well, guess what?  This isn't true of any other product anywhere, nor will
it be true of any product anywhere.  Look at the huge number of different
designs that exist for pencils, or screwdrivers, or coffee cups, or any other
consumer product for which the technical problems have been completely
solved, and the rest is design style.  Linux has a place, Digital UNIX has
a place, OpenVMS has a place, Windows NT has a place, Windows-95 has a
place, MVS has a place, OS/2 has a place, Macintosh has a place, etc.
The Linux place may be growing, and in fact it may grow at the expense of 
some other O/S.  But it won't grow at the expense of *every* other O/S.

As a technical pre-sales person for Digital, I am asked all the time which
O/S I recommend.  The answer is obvious: the three most important things
about buying real estate are location, location and location, and the three
most important things about buying a computer system are applications,
applications and applications.  Find the app which satisfies your needs
(all your needs, not just price, not just performance, not just functionality,
not just support, not just documentation, not just a critical mass of people
who are using the same software so you can share hints and kinks, etc),
and then see what hardware and software platform it is available on.  Then
ask those same questions about the hardware and software platform, and 
make your decision.

There are many places where Linux shines when you look at the whole picture.
And there are many places where it falls down, when you look at the whole
picture.  And that is equally true of every other major O/S out there, or
they wouldn't be major O/S's!  So be passionate in your support for Linux,
but be aware that different customers have different needs...

-- Ken Moreau

    Re: with Microsoft you have a formal, incorporated, financially liable
    entity which stands behind the software, and which has
    demonstrated a strong commitment to find and fix bugs.
    
    Ha, ha. Ha, ha, ha, ha. Ha, ha, ha, ha, ha, ha.
    
    signed,
    
    Still waiting for an ISDN fix about a year later.  I no longer
    spend 20 minutes on hold waiting for Microsoft support to answer,
    though, so I guess I've at least learned something.
    

    re: the gentleman whose wife is installing Linux...
    
    step 1 - buy a book called 'LINUX Configuration & Installation' (2nd.
    Ed) by Patrick Volkerding, Kevin Reichard & Eric F Johnson.
    
    step 2 - Spend a week reading through it, making sure you understand
    everything - ask questions
    
    step 3 - put aside 2 days to actually _do_ it
    
    	I spent 1 half day installing Linux itself, and another half day
    configuring X. Follow the steps one by one in the book, and write
    *everything* you do down.
    
    Also, have documentation available for *every* piece of equipment in
    the system, know what the monitor is capable of (scan rates, etc.),
    know what the video card is capable of. Note down before hand the
    configuration of everything, e.g. extra serial ports, scsi cards, hard
    drive configurations (Cyl/Hd/Sec), ethernet adaptors, irq's used,
    everything!
    
    If all else fails, ask in NLFDC::LINUX-USERS.NOTE
    
    H
    
    p.s. I have no connection whatsoever with the book mentioned above - it
    has been a *very* useful tool insetting up my system, and I can highly
    recommend it to anyone setting up Linux

    re: .58
    
>I guess my only problem with the excitement that you and others have around
>Linux is the idea that it is the obvious choice for every computing need.
>That it can and should replace every other O/S out there.  That it can
>totally satisfy every single customer.

    We're clearly in violent agreement.
    
    I dwell in no land of illusion where Linux reigns supreme in every
    application.  By no means.
    
    Do you dwell in the land of denial that says "Linux is insignificant;
    we should safely ignore it"?  It doesn't sound like it, yet I know many
    who would make such an assertion.
    
    My job, as I perceive it, is not to foster the false notion that Linux
    is THE ultimate O/S; it is to provoke people to consider the fact that 
    Linux is FAR more significant than we (Digital) are willing to believe.
    
    I get REALLY tired of this "niche market" argument.  DIGITAL ONLY DEALS
    IN NICHE MARKETS.  We do NOT own the desktop.  We do NOT own the server
    market.  We do NOT even own the O/S that we've made our "showcase
    offering".  What DO we own?  Some genuinely hot pieces of technology that
    face an uphill selling cycle because they're in markets defined by other
    people's architectures.
    
    The Linux market currently has many more Intel-ish boxes than Alpha
    boxes, but our little band of Linux die-hards have worked so
    efficiently that many people using x86 boxes are looking longingly at
    Alpha and saying, "Boy, if I could just get hold of one of THOSE boxes!"
    
    If we could give Linux the added push of a "stamp of approval" by a
    (still) major computer manufacturer, we could see Linux sweep into
    corporations worldwide while making Alpha/Linux the BENCHMARK in the
    corporate Linux marketplace.  If we wait it out, Linux will certainly
    continue to gain corporate presence, but Alpha will probably lose
    ground as Linux is ported to other platforms and we drop in mindshare.
    
    Our current perception of Linux is frighteningly similar to our
    perception of PCs just a few years ago:
    
    	"It's technically inferior"
    	"It's a niche market too small to worry about"
    	"It will never affect our customers in a significant way"
    	"Our customers prefer our high quality, not this cheap junk!"
    
    We let that wave sweep right over us and now we're fighting to stay
    afloat in a whirlpool that threatens to drown us.  It's time we dared
    to ride atop a wave, instead of just treading water as it thunders over
    us.
    
    -- Russ

    re .59:
    
    I know Microsoft support (if there is any...) is often difficult to
    deal with. However, I participated actively in the Windows 95 beta, and
    at least during that time, the support was excellent. I was contacted
    by the actual engineers doing the work; was sent fixes by email, and
    interim releases (between the builds that were actually pressed as CDs)
    on CD-R's, by courier from Redmond to Munich.
    
    re .61:
    
    �I get REALLY tired of this "niche market" argument.  DIGITAL ONLY DEALS
    �IN NICHE MARKETS.  We do NOT own the desktop.  We do NOT own the server
    �market.  We do NOT even own the O/S that we've made our "showcase
    �offering".  What DO we own?  Some genuinely hot pieces of technology that
    �face an uphill selling cycle because they're in markets defined by other
    �people's architectures.
    
    While it's true we don't own Windows NT, I'd hardly call it a niche
    market. Windows NT server sales have overtaken Unix server sales (not
    even mentioning client/workstation sales). What market share does Linux
    have in the _commercial_ Unix marketplace?
    

    And when we talk market share, can we split it down by Linux variant,
    since there's not just one Linux, but many ?
    
    If anybody in Digital can make money from Linux, fine. Digital
    Semiconductor are well placed to do so, and I wish them luck.  In the
    markets I deal with, some folks use it as (part of) a development
    environment. On non-DIGITAL Intel. But most of the interest I have
    heard from potential Linux/Alpha customers has come to nothing, partly
    because I'm in Europe and it's so much cheaper for folks who are
    serious about Linux to go buy boards in the USA. They don't need local
    support, they are their own support. So why pay DIGITAL Europe prices?
    
    Any tryping mistakes I didn't splot?
    
    regards
    john
    

    What market? The same one that sun and HP have beat us out of...
    the minds of the future. My son's friend is in college and
    he raves about Linux. I watched as Vax customers had the same
    reaction to Sun's and to a lessor extent HP when they were new. 
    For some strange reason they seem to effortlessly dismiss Dec
    in all their dealings.
    
    The site I work at has a space component which goes with whatever
    the latest trends in the colleges are and it pains me to not see 
    Alpha's as a significant part of them. This Linux is a bold
    idea and should get some real funding not just spare time from Digital
    Semiconductor but the Unix people who know good ways to make it work 
    well on ourboxes etc.
    
    Fred

    re:-1
    
    Believe me, the people at Digital Semiconductor who are doing the
    Linux work are Unix people. That is at least one thing you do not
    have to be concerned with. Although a small operation, it is their
    full-time job. Linus Torvalds loves Alpha. Says so in public all the
    time. I have heard this with my own ears. The DS people doing the
    Linux work are not chip designers who are sort of hobbying with this.
    They are experienced long-standing Unix system developers. It's okay.
    Not to worry.
    
    /Charlie (I wrote the early Linux/Alpha "blade" documentation)

    re: .65
    
    I'm THRILLED that Linus Torvalds love Alpha.  It's EXACTLY the position 
    we need to be in!
    
    But, we need to publicly say, "DIGITAL loves Linux (on Alpha)".  Give 
    the current small Linux team enough funding to add a few more highly 
    industrious, highly talented Unix-type engineers & the like to make 
    certain that Linux/Alpha remains a high quality offering within the 
    Linux space.  Have a few of the bigwigs within this company embrace
    Linux (emphasizing Linux/Alpha, of course  8^) in the press.  Bring 
    in some support people focused on Linux/Alpha.
    
    Externally, we should make noise about us being the ONLY major vendor
    to actively promote Linux (and almost no sentence should mention Linux
    without mentioning Alpha; build the mindshare that the CORPORATE Linux
    solution is an Alpha solution!).
    
    Internally, the team should focus on Alpha-specific issues (largely
    kernel and drivers, I would suppose) to make sure we stay on top in
    performance and reliability.  The bulk of Linux development (utilities,
    apps, etc.) should continue as it is now: elsewhere.
    
    We could offer Linux consulting services (and, yes, we had better be
    prepared to do more possible custom work in C or C++; possibly custom
    driver work, etc.).  The Linux NSIS component.
    
    We could encourage Caldera and the growing commercial vendors to cuddle
    up with us.  Distribute a few seed units.  Give them some free
    publicity as they stand next to our VPs during press conferences
    proclaiming the goodness of the Linux/Alpha product offering for the
    corporate world.
    
    
    re: "where's the market?"
    
    This point is critical.  CRITICAL.  The market is building.  It's
    snowballing.  It's growing AT A VERY IMPRESSIVE RATE!  We need to sign
    on NOW -- BEFORE IT GROWS ANY FARTHER ON ITS OWN.  If we wait until the
    numbers are there before we even begin to plan major corporate
    involvement, we will NOT achieve the desired effect!
    
    Why?  If we wait until the market goes full bore on its own accord, our
    ever-so-swift competitors will steal our thunder!  Or, the Intel
    architecture will become the defacto standard for the commercial space
    on the basis of the number of installed units alone.
    
    Like Ken M said earlier:  corporations want some security.  If DIGITAL
    provides that security through statements that we'll stand behind our
    Linux/Alpha products, the boxes that appear in the corporate world will
    quite possibly be Alphas.  If we wait until the corporate market grows
    on its own accord, the platform which delivers the security to
    corporate buyers will be Intel by sheer virtue of the number of
    installed units.  Alpha will probably have a place -- but a small place.
    
    I believe DIGITAL can become _THE CORPORATE LINUX_ provider of choice.
    
    We can help _SHAPE_ this market, not just survive in it.
    
    -- Russ

> demonstrated a strong commitment to find and fix bugs.  With Linux you
> have an ad-hoc, informal, nonexistent from a financial standpoint, and 
          ^^^^^^  ^^^^^^^^  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> completely unaccountable group of constantly changing people who have a
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Errr... Does that sound like something else?  The internet?
(A few years ago)

BTW: I'd also agree that Linux is certainly not the solution
for everything.  Just like BayBank doesn't use the internet to
send bank transations from ATMs.

John

re: .66

>    re: "where's the market?"
    
Okay, so maybe there's a "market".  More to the point, where's the opportunity
for *Digital* to get rich?  Your arguments seem to point mostly to selling
Alpha boards & boxes, and maybe some potential success in services.

If one concedes the opportunities mentioned above, what do we do with the rest
of us who develop software for a living?  TFSO?  Boards & services don't
sound like enough to keep 50,000+ of us going.

(Never mind -- The handwriting's on the wall even without Linux...)

-Mark

    
    
    But what percentage of the Linux market will be willing to pay DIGITAL
    Alpha [premium] prices?  We are unable to convince many commercial 
    accounts to pay a premium for DIGITAL Alpha.  Why would the Linux crowd?
    
    

|    But what percentage of the Linux market will be willing to pay DIGITAL
|    Alpha [premium] prices?  We are unable to convince many commercial 
|    accounts to pay a premium for DIGITAL Alpha.  Why would the Linux crowd?

	Today, they Linux crowd would probably be more happy with refurbished
	old gear, like multias. But if the trend continues, Linux will begin
	to penetrate at the low end in real corporations. It certainly would
	not hurt our argument to stand behind Linux/Alpha in those cases.
	Perhaps we are standing behind it already, but just too far back in
	the shadows.

	Why would the Linux crowd go with Alpha instead of Intel? Could be
	as simple as convenience. The present UNIX market is full of choices.
	Perhaps the Digital choices do not stand out enough in the commercial
	space. The Linux market is not that full yet. Anything serious we
	do could make us stand out alot more, relatively speaking, than in
	the traditional UNIX market.

	>Per

    re: .68
    
>If one concedes the opportunities mentioned above, what do we do with the rest
>of us who develop software for a living?  TFSO?  Boards & services don't
>sound like enough to keep 50,000+ of us going.

    I am not advocating that we toss our current areas of interest and go
    with Linux instead.  I am saying that this can bring us some
    (potentially serious) additional revenue and growth.  It's an
    opportunity to grow new revenue streams and gain new marketshare.
    
>(Never mind -- The handwriting's on the wall even without Linux...)

    Unfortunately, yes, and the writing on the wall reads:
    
    		Windows NT
    
    We've already decided that we can live in a market where we give up the 
    O/S to another vendor.  If we wanted to invest in software, maybe we
    could sell some of our nifty Digital Unix utilities scaled back onto
    Linux for a few bucks (get Walnut Creek to do some retail CDs of the
    stuff).  It might get enough DU flavor out there to get people to
    upgrade to DU in time.
    
    On second thought, never mind.  We don't do software anymore, right?  :^(
    
    -- Russ

>    But what percentage of the Linux market will be willing to pay DIGITAL
>    Alpha [premium] prices?  
    
    I knew someone would raise the issue of Alpha pricing...With the recent
    price decreases Alpha is within $100+ of the latest Intel box (see
    comparison in http://www.digital.com:80/semiconductor/mvi/P010.html).
    That is the good news.  The bad news is that this pricing is available
    for large lots which is contrary to attracting the "masses."  
    
    Another situation that currently exists in favor of going after market
    share (if supply can be met) is that there is a shortage of PPRO chips. 
    Do you take this window of opportunity by setting one price for the small
    as well large and go head to head with the mother-of-all-chip-producers
    or keep current pricing and ignore the small OEMs which is clearly where
    the LINUX market is?
    
    Dan
    

    re: Market share and hardware prices...
    
    one of the main reasons Alpha computers (and we seem to be focussing on
    the 'desktop' at the moment) are so "expensive", compared to, say, a
    Pentium Pro 200 or whatever, is lack of volume.
    
    If, looking at a purely markeing point of view, we (DEC) can _leverage_
    Linux on Alpha to increase the number of desktop boxes shifted, then we
    may be able to reach 'critical mass' where the price of the chips will
    go into free-fall, and the motherboard manufacturing will be comparable
    (? within $US50 ? - I dunno, I can but dream ;') to the Intel boards,
    then an AlphaPC with Linux will take off!
    
    I've noticed on the web that there are Alpha PC's from third parties
    where the prices are within about $1000K of a similar Intel PC - we
    should be able to get this differential down to less than $500.
    
    A great opportunity, but I fear we will squander it.
    
    H

    Digital Semiconductor has been funding the Alpha Linux
    development for the past two years.  They have chosen a path of
    helping the Linux community with the port, rather than doing it
    themselves.  From a marketing standpoint, we have chosen the path
    of helping OEMs, VARs and resellers sell Alpha Linux rather than
    sell it directly ourselves.
    
    To answer some of the questions/statements here:
    
    Linux vs Digital UNIX:  They are complementary.  While there
    might be some customers who would buy Linux rather than Digital
    Unix as a less expensive alternative, I have found that there is
    a much greater uplift of people who know Linux, find out about
    Alpha, then end up buying Digital UNIX (despite its higher cost)
    because of either its features or the fact that it has 5000+
    commercial applications on it.  But without that path, they would
    not know about Alpha, Digital and Digital UNIX.  Bottom line:  I
    have sold more Digital UNIX with Alpha Linux than without.
    
    Linux vs WNT:  I am sure that there are people inside Digital
    that feel we have to put all our eggs in the WNT basket, that we
    can not concentrate on anything other than WNT.  Alpha Linux is
    simply selling incremental Alpha systems to people that would not
    use WNT if it was given to them.  This is either due to
    "religion", the fact that they have an immense investment in UNIX
    and they don't want to re-invest in WNT at this time, or that
    they find that Linux gives them everything they need from WNT at
    a fraction of the price.
    
    If a customer comes up to me and says "I want WNT", I tell them
    we have the best there is, and it runs on Alpha.  But without the
    Alpha Linux project, I would have had to send them to Intel, or
    SPARC, or Motorola.
    
    In fact, a selling point of Alpha Linux is that it tends to run
    on WNT specific Alpha systems, which allows the University sector
    to dual-boot either Linux or WNT.
    
    As to the difficulty of installing Linux vs Microsoft products, a
    lot of the "ease" of installing Microsoft operating systems is
    because someone (a VAR, a reseller, etc.) had originally
    installed a system on the hardware.  You can get the same thing
    for Linux today.  Some vendors pre-install Linux and that makes
    it just as easy to buy it as going down to Circuit City.
    
    With a little more cooperation from hardware vendors, the XFree86
    people (the ones that do the X Window distribution) could put
    every vendor's board and every vendor's monitor information in
    their database, and then the X11 installation would be a snap.
    If there was a really SMART vendor out there, they would put ALL
    their information in the database, and then THEIR hardware would
    be easier to use with Linux (netBSD, FreeBSD, etc.) than any
    other vendors.  Do I hear neural activity?
    
    We should package our own version of Linux:  We made an early
    decision not to do this.  Red Hat, Caldera, Slackware, Yggdrasil,
    Debian, and several more distributions exist.  No one knows who
    will "win"?  Maybe several will continue to be popular.  Why
    should we piss these people off by coming out with our own
    distribution?  In addition, this is another reason not to sell
    Linux itself through our direct sales force.  They would
    (rightly) want to know which one Digital recommends.  By moving
    this out a level to the distributors and resellers, it allows the
    customer and the marketplace to determine the "winner".  Note
    that this does not preclude our sales force from selling
    hardware, but it should be from a order-taking standpoint, and
    let the customer specify the release of Linux they want.
    
    Digital could make even more money with Linux, with just a little
    more support from the various groups:
    
    If the *BUs published information on their web pages about the
    video controllers they used in their systems, the SCSI
    controllers, the scan rates of their monitors, etc. then it would
    be easier for the Linux community to support them, and the
    reseller/integrators to put Linux on our machines (or netBSD, or
    FreeBSD, for that matter).
    
    If the *BUs either kept a current copy of Linux and tried it on
    their new equipment, or made their new hardware available to the
    DS engineering group, this would help us support Linux from the
    very first shipment.
    
    If the corporation would help us optimize the GNU compiler suite
    code generation package to get another 10-15% performance out of
    the EV56 CPUs, that this would not only help Linux, but netBSD,
    FreeBSD and even Digital UNIX sales.  Despite the fact that DEC C
    and DEC Fortran are really great compilers, some people have
    standardized on gcc across all their platforms, and the lack of
    good optimization on Alpha makes it look like our chips are
    slower than they really are
    
    If the multivendor customer services group was willing to provide
    front-end 24x7 telephone support world-wide on a time and
    materials basis for some of these software houses (Red Hat,
    Caldera, Yggdrasil), they might find some additional income from
    these vendors.  Why not a full-service offering?  Perhaps over
    time, but right now Linux is changing so fast, it would probably
    be best to just translate the call, track it, and pass it on to
    the vendors.
    
    Build in greater Linux/Digital UNIX compatibility:  While there
    is a great deal of compatibility there today, a little extra work
    (est. at 12 person months) might allow Linux/Alpha binaries to
    run on Digital UNIX.
    
    A little more work might allow Intel binaries of various types to
    run on Alpha Linux (and perhaps from there to Digital UNIX).
    
    Advertising:  Yes, to have a system that has no license with it,
    that would be "Linux ready", and to make this one of Digital's
    recognized operating systems, would be great.  To be able to have
    Linux show up on our web pages, mentioned by our PR people, etc.
    would be great.  So far our "advertising" has been funded by
    engineering people (and one over-worked marketing person) writing
    technical articles, attending conferences, and helping out our
    resellers.  The money (what little has been spent) was spent VERY
    carefully, and we leverage what we can.  For instance, there was
    a four-part series of TV shows done by PCTV (http://www.pctv.com)
    on UNIX and Linux.  The Alpha was prominent throughout this
    series, estimated to have been seen by three million people.
    Cost to Digital for the series?  NADA.  Nonetheless, a real
    marketing budget (however small) would be nice.
    
    Bottom line, Linux is here.  Either Digital can help it along,
    nurture it on the Alpha and look like heros to that community, or
    we could ignore it and look like a goat.
    

    Russ,
    
    Why don't you, Mr. Hall, and your other mentioned buds put together a
    little business proposal and shop it around to the SBU crowd, starting
    at Debbie Miller and her 70 (oh my!) direct reports.
    
    The thought of our missing another PC-ish market is scarey.  Even if
    it's only 10% of that, we can't afford to miss it.
    
    Bruce Langston
    

    Why do you think Mr. Hall (maddog) hasn't done things along those
    lines already?
    
    The real question is: When will DIGITAL wake up?
    
    I still remember maddog's proposal to get Alpha boards out to EDU,
    publicised by DECUS, to establish a Linux base.  He was right two years
    ago, and he is right today.
    
    (I still remember maddog's advocacy of including a database in ULTRIX
    and still believe he was right.  I just think it will be yet another
    ten years before anyone realises it.  Even VMS has been turned upside
    down by the sale of Rdb to CAI).
    
    fjp
    

	RDB was sold to Oracle, not CAI.

	Folks, it's not a Linux problem we have here. It's a DIGITAL
	problem. We can't see an opportunity when it is staring us in
	the face. The beancounters run this company and we have a
	marketing organization that is not allowed to market.

	Fix DIGITAL and the rest will follow. Alas, I'm afraid it's
	becoming too late.

							mike

    in 10 yrs, what will be the market share of linux compared to other
    variants of Unix?
    
    what will be the market share of Digital Unix at that same time?
    
    can we afford to ignore either?
    
    why is long-term thinking so difficult - and so infrequently donee in
    Digital (nee DEC) these days?

    
    
    Maybe the Linux champions can post here [even the outline of] a Business
    Plan that represents what actions and investments Digital should make  - 
    with return on investment dollars?  
    
    The Linux market, which strives to be most "open" and you get
    Linux of $39.95 in the back of a book, hardware/CPU is NOT a commodity 
    item???  
    
    Digital is not winning in commodity computing markets. We are niche.
    If we want to fight and maybe win in a commodity market - we should 
    choose a market with big prize (WNT) that matches our strengths and 
    focus all our efforts on winning.  If we fight and win Linux - what 
    will we have won?  And what other opportunities will have been missed
    in the process?
    
     
       
    
    
    
    
      
     

    Since I'm not in a place to know how much we are currently spending on
    the Linux/Alpha effort, I won't comment on dollars for future funding
    -- I have nothing to go by.

    I can, however, take a stab at your concerns:

>    Digital is not winning in commodity computing markets. We are niche.

    Correct.  Digital is not winning in current commodity markets because:

    	a) the straight Windows/Intel market is flooded with low cost
    	   alternatives; we don't stand out.

    	b) our WNT/Alpha program is constantly fighting the uphill battle
    	   on not being x86 compatible.  Even with our blazing Alpha
    	   speeds, the Windows market sees us as incompatible -- and that
    	   means a real tough sell.

>    If we want to fight and maybe win in a commodity market - we should 
>    choose a market with big prize (WNT) that matches our strengths and 
>    focus all our efforts on winning.  

    Great!  We'll just bowl over Intel, AMD, and Cyrix... NOT!

    You want to WIN the WNT market?  Get Alpha prices BELOW Intel's!  For
    that, you need serious volumes.  Better start shipping BOATLOADS -- and
    I MEAN BOATLOADS -- of Alpha chips throughout the world.  How you gonna
    do that?  Try winning a growing market that might start eating them up
    without needing the uphill sell.

    Look at the potential Linux/Alpha "commodity" market.  What are the
    rules?  Must run Linux.  Is x86 compatibility much of an issue?  No.

    In fact, much of the current Linux community looks at us and drools
    already.  But the current market is only now expanding to the corporate
    world.  If we step up to the microphone and say "Linux is here and
    DIGITAL Alpha is the CORPORATE LINUX PLATFORM", we can make a serious
    bid to WIN the corporate Linux market.  We have a level playing field
    and we can finally compete based on our strength: breathtaking speed
    (and professional consulting and hardware maintenance services, if we 
    have the nerve).

    >If we fight and win Linux - what 
>    will we have won?  And what other opportunities will have been missed
>    in the process?

    What will we win?

    	1) Money.  Cold, hard cash.  Why should a Linux/Alpha sale be bad,
    		when a WNT/Alpha sale is good?  There is still real cash
    		to be made, if we have a mind to make it.

    	2) Volume.  Start spitting out Alpha chips to feed the hunger of
    		the Linux market.  Maybe we can actually start dropping
    		prices and FINALLY go head-to-head for MAINSTREAM WNT.

    	3) Presense.  Digital is now perceived by many as the next Prime,
    		the next Wang.  Oh joy.  Bet your business on us?  But if
    		we grab hold of a market that already has an excellent
    		growth rate in education and is clearly making its way
    		into the corporate world, we can be seen as a SERIOUS
    		VENDOR OF NEW TECHNOLOGY.  Read maddog's reply again:
    		He sells more Digital Unix because we support Linux/Alpha
    		(in our current low key fashion).
    
    	4) Future.  Remember: the O/S used by today's students often shows
    		up in tomorrow's corporations.  Let the college students
    		use Linux/Intel.  When they start making business
    		decisions, they'll want Linux.  What hardware?  "Hey, I
    		remember seeing/hearing about that Alpha stuff in college!
    		Man, it runs Linux FAST!  Let's get one of those!  Oh,
    		what's this, they've got another thing called Digital Unix
    		which is supposed to be a lot like Linux, but even better? 
    		Well, let's look at that, too, then"
    
    What will we miss?  I can't say that we'd miss much.  I don't here
    anyone saying this is a billion dollar investment with people being
    pulled off their tasks across the globe.  Some modest engineering
    investments (see maddog's reply) and some serious, high-level press
    conferences and the like could do quite a bit to start with.

    If people will look at us who have never looked at us before,
    Linux/Alpha may prove to be the best corporate advertising we have
    done in years.  And we can turn some serious profit from the effort as
    well.
    
    -- Russ

    re: .77
    
    Oracle, not CAI.  Brain fuzz, because CAI ended up with Ingres, the
    DBMS we were bundling with ULTRIX.
    
    Sorry for any confusion.

  RE: <<< Note 5120.80 by NEWVAX::PAVLICEK "Linux: the PC O/S that isn't PC" >>>
 
    This bears repeating, over and over (like a mantra of sorts?)
    
    >  	4) Future.  Remember: the O/S used by today's students often shows
    >		up in tomorrow's corporations.  Let the college students
    >		use Linux/Intel.  When they start making business
    >		decisions, they'll want Linux.  What hardware?  "Hey, I
    >		remember seeing/hearing about that Alpha stuff in college!
    >		Man, it runs Linux FAST!  Let's get one of those!  Oh,
    >		what's this, they've got another thing called Digital Unix
    >		which is supposed to be a lot like Linux, but even better? 
    >		Well, let's look at that, too, then"
    >
    
    We seem to be in corporate mindset that completely ignores the most
    obvious of facts. You reap what you sow. If you do not plant the seeds,
    NOTHING WILL GROW! I've seen it over and over, the bright young eager
    go-getters working their way up the corporate food-chain, providing
    input on what to buy to implement <insert project name here>.
    
    We lost the workstation war to Sun on BSY2 due to one (ONE!) system
    type who had the ear of the planners, how had worked with Sun boxes and
    Other flavors of Unix in college. We PROVED better performance and
    price, and still lost. 
    
    Here is yet another oportunity to strike while the iron is hot and I
    fear that the beancounting-stovepipes will turn a blind eye to it.
    
    .mike.

Copied from comp.sys.dec, just for interest. (em86 is to Linux on Alpha as FX!32 
is to Windows NT on Alpha.)

PJDM

Path: 
pa.dec.com!news1.digital.com!su-news-hub1.bbnplanet.com!cam-news-hub1.bbnplanet.
com!news.bbnplanet.com!howland.erols.net!europa.clark.net!newsfeeds.sol.net!nntp
.uio.no!uninett.no!nntp.uib.no!postmaster
From: Ketil Z Malde <[email protected]>
Newsgroups: comp.sys.dec
Subject: Digital Linux?
Date: 09 Apr 1997 04:30:18 -0400
Organization: University of Bergen, Norway
Lines: 33
Sender: ketil@triton
Message-ID: <[email protected]>
References: <[email protected]>
NNTP-Posting-Host: gatekeeper.imr.no
X-Newsreader: Gnus v5.4.37/XEmacs 19.15

[email protected] (David Rudder) writes:

> I checked out em86 last night, and I have to say that I am quite
> impressed.

Cool.  You know, every once in a while, we the people are reassured in
our faith that - NT nonwithstanding - that Digital is basically a good
company.  First they port Linux to the coolest hardware that a limited
amount of money can buy, then they port their proprietary X-server for
TGA, now they give us EM86.

Few other companies seem to do that -- in fact, many hardware-only
(-mostly?) manufacturers even refuse to give out specs to their hardware
to the Linux people.

Clearly, giving out code like X/TGA and EM86 requires some kind of
permission from above -- this is, one would imagine, proprietary code
that Digital has a substantial investment in developing.

However, I see very little about Linux on Digital's www-site -- in fact,
I find nothing.  What I would ask somebody from Digital is, how official
is Digital's support for Linux?  I'd imagine (at least taking this
newsgroup into account) that Linux runs on a large share of the
low-endish Alphas out there -- why not a www.linux.digital.com or some
such?

I'm not blaming Digital for not supporting Linux enough, far from it.
I'm just curious as to how Digital feels Linux fits into their
organization (and when can we please have Digital's compiler technology? :-)

~kzm
-- 
If I haven't seen further, it is by standing in the footprints of giants

    My management has told me that their management says "we won't put our
    compilers on Linux".  Too bad.
    
    					Steve

    So Steve,
    
    I guess this was after a detailed assessment of the marketplace
    segmentation and potential? Right?
    
    - John

    
    	You must be a new hire or something.
    
    
    
    	8^)
    

    
    Yes I am, November 1985. But dont worry, I'm having "Introduction to
    Digital" training real soon.
    
    I've stopped reading Dilbert, we've moved THE CORPORATION to a higher 
    level of farce than Dilbert could ever sustain...
    
    - John
    
    p.s: memo to the branding committee, the button marked "caps lock" lets
    you do lower AND upper case.
    

RE: .87

	Settle down Asok...

	:)

							mike

    
    Um, wouldn't giving them our compiler technology require putting
    GEM under the GNU General Public License?
    
    A man who rarely understands licensing,
    DFW

    re: .89
    
    I don't know, but I suspect Steve's reply was to the notion of SELLING
    our compilers on Linux.  Just like you can buy WordPerfect for Linux,
    would could sell our FORTRAN for Linux, etc.
    
    Now, I can't say for sure that the person on the Internet post wasn't
    asking for a donation...  8^}
    
    -- Russ

What the company has done is make alpha compiler know how available
through the on-line "compiler cookbook".  It is rich in technical information.
Try "compiler cookbook" in altavista.  Its pretty good!

Jon Hall's vision is the way to go. Stay with freeware for entry
level linux.  If they get really serious and need to scale up then get
DUNIX with the best alpha compilers.  

Is there a problem here? 

 

    I don't think that making and selling SW for Linux requires it be
    released under the GPL - as long as you don't re-use any code that had
    previously been issued under the GPL.  

I thought Linux was binary compatible with Digital Unix, so you should 
simply be able to copy the Digital Unix compiler executable to a Linux 
system and run them.  Of course, this ignores the licensing issue.

Management is probably worried that to be successful with Linux requires a
high volume, low price, which conflict with Digital's strategy of low 
volume, high price.


						-Paul

The compiler is only a small piece of the picture.  What about the RTL?  Does
it call any DU interfaces not present in Linux?  What about debugging?  What
about documentation and support?

				Steve