Conference 7.286::digital

    Isn't that just what has been tried, repeatedly, with UNIX?  Remember
    OSF/1? I see that it's being tried again with SCO as a base.  The problem 
    with this is that "design by committee" is too slow and conservative.
    The only real advances made in OS design, features and usability have
    been brought forward by vendors pushing their own variant (or, to a
    stronger extent, their own "proprietary" OS.)   Digital UNIX is only
    now getting around to features, such as clusters, that VMS has had for
    ten years.  When we don't control the development, we can't push the
    state of the art, and customers will turn to vendors who can.
    
    				Steve

And the bottom line is differentiation. We can't gain market share
without some percieved "value added". And every vendor wants to
exploit the strengths of their own hardware. If SUN adds a feature
we have to add a feature and so on.

In some ways computers seem like cars. Most all of them get the job
done in a reasonable manner. So, in order to convince someone to buy
our model we add some chrome and more horsepower. The problem has 
become that computers no longer have even a year's cycle time between
new models. liesl

p.s. And I've personally been much happier since they gave us back 
little disk lights. :*) 

  Steve's right, of course, when he cites OSF as the latest example
  of the failure of "design by committee". The once-and-Futurebus
  is another recent example.

  But it could be reasonably argued that Windows/NT is an example
  of software that was deliberately designed to run on all machines.
  But it was essentially the brainchild of one pig-headed person,
  not a committee. It was Dave's way or the highway, as is well-
  documented in the book "Showstoppers".

  I think most of the innovation in our business remains the result
  of one or a few individuals with a bright idea or the exquisite
  timing to recycle an old idea at just the right time.

                                   Atlant

    I received a flyer from those nice people at HP today which headlined,
    
    '64-bit computing, High performance or hype?'
    We are on the verge of the 64-bit computing environment. An environment
    that promises the power to run the ever more resource intensive,
    business critical applications needed in today's competitive world
    
    (The flyer then introduces a seminar offering about 'A NEED FOR
    STANDARDS - Open systems based on UNIX')
    
    'Mike Lambert, VP and Chief Technical officer, The Open Group, will
    examine the issue (of many Unix flavours) in detail...'  They then want
    to tell the seminar why the HP/SCO and HP/Intel initiatives are both
    leading to an environment which will provide advanced system
    management, clustering, advanced software management etc. etc. ad
    nauseaum.
    
    Isn't it good that HP have the industry's interests at heart and
    offering a preparation 'for an evolutionary transition to a 64-bit
    world'?  All based on universal Unix.
    
    Calvin
    
    

    I registered for that, Calvin. A few days later a very polite lady rang
    and said I could not attend.
    
    Nice to know H-P are in control.
    
    Andrew

    
    It exists, but doesn't yet run on _all_ machines.
    And yet, it isn't controlled by a single company.
    
    The single most influential developer of this OS is an avid Alpha
    fan, and will say so in public.
    
    Conservative estimates indicate this OS is running on more than
    500,000 machines currently, and growing.
    
    It's a freely available 64-bit OS called
    
    				L I N U X
                           

    Oh yes, a new holy grail for the Unix crowd. L I N U X !!! Gag and
    retch!!! The true value of Linux is if you are prresenting a UNIX
    solution and someone mentions a competitor you can bring up LINUX and
    the true unix heads fall into immediate arguement over "the true, good,
    one and pure operating system is _______". It's great. The pony tails
    and sandals fly. Better than bringing up Hinduism at a Baptist
    convention.

    Repeat after me...  If it sells HW...  fast HW...  Alpha HW... then IT
    IS A GOOD THING...  Sandals and ponytails are irrelevant...  Sheesh!! :-)
                                                                          

    Before you dismiss Linux totally, did you know that there is an
    implementation of Linux which is:
    
     o  running on Alpha already
     o  US Government FIPS Compliant already
     o  X/OPEN XPG certified already
     o  heading for Single Unix Specification branding
     o  binary compatible with Digital Unix executables already
     o  not orderable in nice clean packaged form today (why!!!!!!!!!)
    
    I'm saying nothing here other than "make sure you are fully informed
    about Linux before you dismiss it". Open systems, open minds, etc.
    
    regards
    john

>      o  not orderable in nice clean packaged form today (why!!!!!!!!!)
    
  It appears to be orderable on cd-rom from red hat

http://www.redhat.com/redhat/news_and_info/PR/redhat-3.0.3.html


                           Mark

    Sorry for the toe trodding. Have to watch those open shoes. (a joke,
    meant to be humorous in a non-threating comradely way) I was uninformed
    that a true open-system had been created. Imagine being able to buy an
    application in LINUX and being able to run it on any platform running
    any version of LINUX. I hadn't heard. I thought there was a LINUX
    kernel and hundreds of variations of system services and executables
    added by as many different developers. I stand corrected, being careful
    where I stand of course.

  So I can take a LINUX binary compiled for a, say, '486 and run it
  on my Alpha? Or a PowerPC binary and run it on my MIPS? Or Alpha?
  Or 'X86?

  Sorry, but I think there's more to life than "the operating system
  runs on all the platforms". Even 'NT isn't able to completely solve
  the problem of "n architectures -- n binaries", but things like
  high-powered dynamic-recompilation emulators go a long ways towards
  solving the problem and those high-powered tools *ARE* starting to
  be available on 'NT -- Are they also available on LINUX?

                                   Atlant

    Red Hat commercial Linux for Alpha is available now (I got the CD last
    week) but a.f.a.i.k. their package isn't FIPS compliant, XPG certified
    etc. Lasermoon's is (but isn't neatly packaged yet!!!!). 
    
    Free software is worth every penny. Just like free advice.
    
    Digital Unix, and every other Unix known to me, consists of a kernel
    and a bunch of code strung together by a network of unrelated
    developers. Just like Linus and his mates. Except Linus gives you the
    source code so if you feel creative you can play.
    
    Meanwhile, as regards single binary for every platform: Someone (the
    OSF?) was researching exactly that, using a thing called Architecture
    Neutral Distribution Format. It's been shown at trade shows and so on
    and progress meetings have doubtless clocked up thousands of Air Miles.
    There are more commercial issues to sort than technical ones. I'm
    expecting a product to be available the week after Digital announces a
    freeze on VP recruitment. (UCSD Pascal, anyone? Multiplatform binaries
    circa 1980...)
    
    regards
    john

Is there some reason that the world needs yavu?
  (yet another version of unix)  

Linux may be pure and wonderful, but is it truly what
the world needs?  

re:  - a few.  something like "if it sells hardware, who
cares"  

Some people would say that software is what makes 
hardware useful.  If that's true, then selling software
becomes important (more profitable, too), and hardware
sales  a less interesting event....

    .15> Some people would say that software is what makes 
    hardware useful.  If that's true, then selling software
    becomes important (more profitable, too), and hardware
    sales  a less interesting event....
    
    Amen Brother!!!
    
    /s/Dan&iSbu
    

RE: LINUX -

Yes, it is a *BEAR*, because it real UNIX, not DOS (or even CP/M).  Why do I
run it?  Because I need to know more about UNIX, but this company does not 
help field folks learn about products beforehand (same problem as S/W 
developers, but I believe I am even further back in the line).  I also happen
to have the minimal H/W config for it (386sx-16, 5MB RAM, 170MB disk, and
a single-speed CD-ROM), BECAUSE I had to give my 486DX66 to my eldest
(Win'95 on the 386sx?  I wouldn't EEEVEN try it).

Things to Ponder on the LINUX FRONT
-----------------------------------
1. [The World's Operating System]
   Large amounts of folks in this world now have access to PC's, and no real
   interest/access to MS software, and lack of $$$ to buy the high-end. 
   (Is it possible that by making it harder to compete with them in the WINTEL
   world, MS and INTEL have actually made it more likely that others around
   the world will just bypass them for something cheaper?)

2. [A 4-Cylinder O/S, not a "V12"]
   LINUX can run on less than the latest H/W:  It is not fast on the 386sx,
   BUT IT RUNS LOTS BETTER THAN WINDOWS 95 WOULD!!  (This also works with 1.)
   Much of the PC base will be 8MB 486xx's for the next several years.  LINUX
   works with this much better than WIN'95, in terms of performance.

3. [Remember: Even Microsoft Thought Nothing of the WWW]
   The Internet makes LINUX evolution (as well as S/W for LINUX) a distributed
   task, with a rich set of communities that can address specific needs.
   (What can these communities do with the resources currently available?  
   Remember where Internet Phone, cheap video-conferencing, HTML,
   Netscape, DOOM (I & II), JAVA, and so much else has come from in the
   past few years).

4. [Why Wait for SCO/HP/SUN/DG/NCR/IBM/LSMFT?]
   LINUX is 32-bit and 64-bit TODAY!!  What other UNIX can say the same
   thing?  If the SCO stuff is really 18 months away, then LINUX has the
   same advantage that Digital UNIX has in terms of lead-time.

5. [Used for Business and Pleasure]
   I belong to a LINUX group.  While many members do look rather shaggy and
   late-60's, others are as professional as the rest of us.  SEVERAL ARE
   SYSTEMS CONSULTANTS THAT ARE USING LINUX TO DELIVER SMALL SOLUTIONS
   TO CUSTOMERS FOR A LOWER COST THAN USING SCO OR UNIXWARE.  THIS MAY BE 
   MOST IMPORTANT OF ALL - LINUX RAISES THE BAR IN TERMS OF WHAT IS THE
   MINIMUM CAPABILITY AVAILABLE FOR THE LOWEST COST POSSIBLE, WHILE 
   SIMULTANEOUSLY MAKING IT MORE DIFFICULT TO COMPETE with something 
   that is just about free.  It took some time even for PC-DOS to catch
   on, but what has happened to CP/M-86, PRO-DOS, and all those other 
   early 80's microcomputer OS'?  Only the MAC-OS remains, because it
   was more than PC-DOS.

6. [In the Knowledge Age, WHAT REALLY COUNTS ANYWAY?]
   What about all the UNIX "capital" that exists in people's heads?  LINUX
   makes anyone capable of building on the UNIX foundation with next to
   no real $$$ investment, except their own time and effort.

People once said bad things about the PC, but I ignored all those comments, 
and was never unhappy I did.  I believe that LINUX may not be for everyone, 
but those that make the effort to dig into it will never regret it.  And
the crudeness, toy-like nature of it may be only part of the future it holds.
Remember who said PC's were toys.  HOW IMPORTANT MIGHT IT BE FOR THE WORLD TO 
HAVE AT LEAST ONE COMMON H/W ARCHITECTURE (the PC one), AS WELL AS ONE COMMON 
O/S ARCHITECTURE (the LINUX one)??!!

Charles

    re .14:
    
    �(UCSD Pascal, anyone? Multiplatform binaries
    �circa 1980...)
    
    It's called Java nowadays...
    
    ;-)

	Firstly Linux is _not_ a toy, it is a serious OS with some 
	serious applications.   A lot of universities run their
	networking from it, and an oncology unit is running it also.
	The fact that it was "developed by committee" is interesting
	as is the fact that most people developing it never met.  These
	facts are also irrelevent just like which office the person that
	designed my car sat in.

	Secondly, Linux has it's place, as does all OSs.   They are
	loved, loathed and used for a variety of reasons.  Linux's
	place is in the academic world with not much penetration into
	the industrial space.   Although there are bundled Linux applications
	used commercially and I'd almost guarentee to find Linux someplace
	in an organisation ('cept maybe Microsoft).  Pretty much all of the 
	alphas that we are selling to run Linux are being sold into 
	Universities right now.   That's fine.  Strangely enough our 
	competition here is exactly the same as for Windows NT - Intel and
	Pentium Pro.  Our problems are pretty much the same too - we're
	too expensive and we're loosing our performance edge generation by
	generation.

	Thirdly, historically Unix has engendered pretty powerful feelings
	with various factions not working together.   Witness the failure
	of OSF.   If Unix is dead, why did our workstation sales increase
	so much?   I also believe that if we abandon Unix on the desktop
	then we're pretty much abandoning Alpha on the desktop too.

	Fourthly, as Charlie Christ said recently in Hudson, I don't care
	which operating systems run on our chips so long as we sell the 
	chips.  It's as pointless being an NT bigot as it is being a Unix
	bigot.

	Dave

	ps I don't wear sandels and I don't have a ponytail

could someone shed some light as to where FreeBSD fits into all this?  I don't
know a whole lot about it other than it's free and pretty solid.  does it just
run on INTEL and that's why there is not much mention of it?  I know of several
companies who actually use it as web servers and are quite happy with it.

-mark

There's FreeBSD and NetBSD.  FreeBSD has focused on the Intel machines which
NetBSD has focused on platform portability.  Both are robust stable platforms.
NetBSD has ports available for Alpha, DEC MIPS, and VAX systems as well as 
DEC Intel based PCs.  FreeBSD only runs on the Intel platforms.

Both Linux and *BSD are decent UNIX-like operating systems.  I happen to favor
the BSD based operating systems but that's a personal preference (I didn't
like the Linux kernel where I looked at its sources nor do I like the GPL
which Linux is afflicted with).  Choose whichever one is easier for you to use
or get support for.

I'm sure someone who went to see Linus at UNH will let me know
if I got this straight or not (I was a-way in the back).

Linus said that while he admired how clean the alpha architecture
is, he couldn't recommend that most folks get an alpha box. :(  I
was suprised to hear him say this, but it seems his opinion is
that the pentium pro boxes provide more bang for the buck than
the cheap-ish alphas.

While I don't necessarily agree with his recommendations, I do 
respect his opinion.  (And I hope we can get enough sales to lower 
our prices)

I was in school when Linux first came about (in comp.os.minix?)
and followed it since.  Until last year I thought it could only
be a tinker's OS.  But in '95 I saw personally saw two companies
using it for buisiness, one more so far in '96.  

One OS?  Probably not.  There are a lot of variables.  How about
one motor for all cars?  Or one standard type of heating/cooling
system for all homes?  Etc...

    I was at the UNH meeting, and have known Linus for over two years.
    He always praises the Alpha architecture as being "clean" and "modern",
    and talks about the speed, even when I am not listening.
    
    Unfortunately, with the advent of the PRO, the price/performance of
    Alpha systems in the 32-bit space vs Intel is questionable.  We charge
    so much for our motherboards (up to $3K street price) that it makes it
    hard to advocate Alpha over Intel.  With a $3K street price for the
    board, the system price climbs to $7K or more.  Then we do silly things
    like pricing the NT systems to compete with Intel (marginally) while we
    price the UNIX systems to compete with HP, Sun and IBM (marginally),
    instead of going for volume penetration.
    
    With Linux, I hope to price low-end "generic" systems to get Alpha out
    to Universities, TOEM projects, etc. while allowing our erstwhile
    management to continue to price Digital UNIX workstations high to get
    that "money left on the table", but fail to make it gain market share at a
    reasonable rate.  I am working with APS, E&RT and even low-end servers
    to do this.
    
    I do not have the time to argue through notes whether Linux on Alpha
    is good or not.  I know that the market is growing, we are selling
    Alphas running Linux, there is interest in it and this interest
    generates opportunities to talk to people about Digital, Alpha and even
    Digital UNIX that otherwise do not know Digital exists.
    
    My hair is too curly to allow it to be tied into a pony-tail, and I
    don't wear sandals, I wear DEC(k) shoes.
               

    LINUX is serious because some Universities use it (.19)? My experience
    with universities is their choice of OS isn't influenced by the problem
    of support difficulty. Downtime? Uh-is that like a uh problem? Ever
    hear of anyone at a university ever being FIRED?
    
    great topic but I still haven't resolved my CP/M vs DOS angst yet
    
    Mike (closet UNIX fan ,once ULTRIX support, former ponytail
    wearer)Wilson 

	Sure Universities care about support - they have to function
	as deliverers of service etc.   I've been told more than once
	that Linux takes less support than a lot of commercial OSs.
	Anyhow, if you're interested in Linux on Alpha, see
	
	http://www.azstarnet.com/~axplinux

	for more information.   No doubt Matt Thomas will post the
	equivalent for NetBSD.

	Dave

    I have two kids that attended Worcester Polytechnic Institute in
    Worcester Massachusetts and they introduced me to Linux.
    
    It seems that WPI is a totally wired campus.  Every room (dorm, office,
    ...) has a LAN drop.  The majority of the students run Linux because it
    provides very good networking capabilities for very few bucks.  So, you
    can do your CS lab in your dorm, and when finished, print it at the
    nearest avaiable printer.  Also, you can get some really nifty MUDs.
    But in any case, the students that are graduating are programmed to use
    Linux.
    
    The above point in important.  Back in the 1970's Digital did a very
    good job of selling into the University market.  Starting in 1975,
    legions of student were coming out of these universities who were
    oriented towards interactive computing.  By the late 1970s and early
    1980s these students were now in positions to recoment hardware
    purchases and guess what they recomended - DEC equipment.
    
    The same effect was used by ATT with Unix.  Basically, Universities
    could get Unix for nothing.  This led to a wholesale adoption of Unix
    by universities and in turn primited the Unix pump outside of the
    Univeristy environment.
    
    My guess is that we will hear more about Linux if the current trends
    continue.
    
    My two cents
    
    -David
    
    
    

    
    Re: .26
    
    As a recent graduate of WPI, I want to toss in another reason students
    are picking linux -> stability.  Once you get linux installed on your
    system, you don't have to worry about not being able to get into the
    network because PC-Xware or other Windows/PC based software doesn't
    work.  Especially when it comes time for the term to end, installation
    and upgrade issues would rear their head in the Windows configurations,
    while the linux users where plugging along, getting work done.
    
    Nat
    WPI Class of '95
    

    
    And lets not forget the WINE project.  Windows API on Linux.
    Free.  Hard not to like a 64-bit OS with multiprocessing,
    multiuser, X-Windows, Windows (coming soon), and global
    support (just ask, someone will answer).

WINE does however require an i386 (or later) to run.  It also runs on other
operating systems other than Linux such as FreeBSD.

    ... and if it's really new you may want to let it age a bit before
    trying it ...

    re .26
    
    Exactly. What made DEC big was fitting into university lab budgets. A
    generation came out knowing what 12737 was. They told their new bosses
    that ..."well I'm sure I can do it if you buy a DEC xxxxx".
    
    Now, how do we do it again?
    
    r

    re .26, .31 - 
    
    Second the "exactly." We are paying the price now for arrogantly
    assuming that colleges and universities were just "customers" and did
    nothing out of the ordinary to seed the places with dec_stuff. A 20%
    discount just doesn't cut it.
    
    Why is Unix so popular these days? Cheap/free to colleges in the past. 
    
    Ask Altavista about "sunsite.unc.edu", that's only one small example of
    us missing the boat again. There has not been a significant digital
    effort to match it. How do I know? My brother-in-law manages
    that site, and hasn't seen any serious effort by digital in years. He
    gets hot new Sun boxes to play with on a regular basis. Guess who *he*
    recommends when asked about new hardware...
    
    Plan for the future, plant hardware and software at the the college
    level (and/or before) and reap the rewards years later.It ain't rocket
    science.
    
    .mike.

>    Why is Unix so popular these days? Cheap/free to colleges in the past. 

dunno about that bit; it's almost 20 years since Unix was free or cheap (not 
counting the sudden flood of Linux or FreeBSD systems in the last 3 or 4 years)

Chris.

    I think its more like ten years.  Back in 1985, Universities could
    still get BSD 4.x for almost nothing.
    
    Back in 1979/1980, a Unix license for commercial enterprises was not
    that cheap.  I think that the going rate back then was in the five
    digits for a single machine license.  If you bought in bulk, the unit
    price went down to the hundreds of dollars per license.

    
    
    .. and the people who were in college in the Mid eighties, using
    cheap/free OSs , grew up with that kind of OS. Now they hold bedgetary
    power, and as we all know, most senior decisions are not based on
    business sense but on individual's perception.
    
    
    
    
    
    

    Granted, without software hardware would be furniture.....without hardware
    software would be hammer..chisel...stone.