Linux-Advocacy Digest #842, Volume #26 Fri, 2 Jun 00 21:13:04 EDT
Contents:
Re: The Mainframe VS the PC. (R.E.Ballard ( Rex Ballard ))
----------------------------------------------------------------------------
From: R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: The Mainframe VS the PC.
Date: Sat, 03 Jun 2000 00:07:15 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> r.e.ballard <[EMAIL PROTECTED]> wrote:
>
> >For those who are under 40, a bit of a history refresher
> >is in order.
>
> May I request a refresh of your refresher?
Sure, why not. I've been in the thick of the PC revolution since 1976
and have lived through most if it. I purchased one of the first copies
of Micro-soft Basic (on punched paper tape) for $25. I even applied
for a job at Microsoft (the first time) in 1982, sending them code and
everything.
On the other hand, there are so many little details, and so many
articles, experiences, meetings, conferences, and dialogues, that
it's possible that I might have forgotten a few things.
> >Actually, Microsoft had very little to do with the cost
> >of the PC. There were 5 factors that made PC's cheap.
>
> ( with MS helping with a possible sixth below... )
We'll discuss that later.
> >The PC was manufactured by IBM, this resulted in a sudden
> >market acceptance by a large body of corporate users who had
> >previously considered PCs to be toys or hobby equipment.
> >With the blessing of "Big Blue", more machines were sold,
> >which made it possible to spread development costs across
> >hundreds of users.
>
> IBM entry gave the "microcomputers" of the time more legitimacy as
> business tools. In particular those that could run programs like
> VisiCalc. IBM's entry also helped solidify the "industry standard"
> in further identifying what included components separated a
> brand name PC from a hobby-kit.
Actually, CP/M was a pretty well established standard. The biggest
problem was that every PC maker had a different standard for their 5
1/4 inch floppy format. The 8" floppies were pretty standard, but
everybody wanted to pack as many bits as they could into the 5 inch
floppy. In some cases, there was no hard drive and minimal networking
that was quite slow (Corvus Constellation, Televideo, and Net-ware).
Drive sharing consisted of passing every sector, across the net.
To open a file, you'd have to pass the directory tracks back and forth.
Actually, Visicalc was available for the Apple and for CP/M systems.
In fact, IBM did everything they could to encourage programmers to
write, or rewrite their programs so that they would "exploit the
technology of the PC hardware". Part of the problem was that MS-DOS
was so minimally functional that people would jump into the
BASIC-IN-ROM to do things like graphics. Display graphics were
painfully slow on a 4 meghertz 8088 processor. The system call itself
took almost 200 microseconds per call.
> >Second, CP/M had established a huge market lead and the
> >original source code for QDOS was based on this same code-base.
> >The issue was settled between Seattle Computer Company and
> >Digital Research many years ago, but Gary Kildall was the
> >original author of what eventually became MS-DOS.
>
> Do you have a cite for that? I have never seen anything that
> absolutely cleared up that issue, only interviews with the
> parties involved where opinions were expressed, but nothing
> concrete either way was presented. Also, QDOS (and hence
> PC/MS-DOS) was available for the 8088 CPU long before
> CP/M-86 was ready, so one could argue that Tim Patterson
> should at least be credited with porting the code earlier
> than Kildall.
Actually, it was most recently published in the briefs filed
by Caldera in the Antitrust suit they settled last year.
Unfortunately the transcript was sealed and the official
briefs are no longer on the web site. It was a secondary
fact. A material fact, but not a strategic part of the case.
Microsoft made no attempt to contest it either.
About 6 months after all of the DRI claims had expired or were
fulfilled, Microsoft paid the remaining SCC directors a total
of about $1 million in cash and hired one of the directors to
a senior Microsoft position.
> >Many hobbiests and computer enthusiasts, most of whom were
> >just graduating from College around 1980 had been developing
> >software on the Apple ][, the TRS-80, and the Commodore Pet,
> >each of which ran versions of Microsfot BASIC. Many of these
> >budding programmers learned to program in assembler or BASIC
> >for CP/M to produce better applications. While Microsoft did
> >write the versions of BASIC used on the TRS-80 and the
> >Commodore PET, there were numerous versions of BASIC in use,
> >and BASIC itself was a product of DEC.
>
> I thought BASIC was created at Dartmouth as a teaching aid.
Correct. DEC was one of the first companies to use BASIC as a
commercial product for business systems.
> >According to a "Gates Fable" he fished
> > the source coded to BASIC
> >out of a dumpster, studied it very
> > carefully, and ported it to the 8080
> >about a year later.
>
> I've never heard that tale...
> According to histories of Microsoft and
> of Gates; He and Paul Allen had been familiar with BASIC for years.
> (He may even have been introduced to it while he and Allen were
> still high-school mates). In fact, this interview:
> http://www.si.edu/resource/tours/comphist/gates.htm#tc4
> seems to imply that as well.
Quoted from your citation: (typos and all).
The BASIC they did which we got ahold of the BASIC source code and
enhanced t quite a bit was great. They had AID and
FORTRAN. They didn't have a very good COBOL until a little bit later
on. But it was a very solid time-sharing system. Lots of
great experimentation going on with it.
> Is this "Gates Fable" something you heard from him, or something
> that was passed more by word-or-mouth?
It's right in the URL you gave. 5 paragraphs below the bookmark.
This doesn't say he fished it out of the dumpster, but the first
I'd heard of it was in 1992, just before Microsoft asked DEC to
help out with Windows NT and promised to support the ALPHA.
> >Many companies had been established using CP/M as the standard
> >operating system. This meant that there were a number of
> >commercial-grade applications such as the original DBASE,
> >Lotus, and WordStar that were available and easily ported
> >from CP/M to MS-DOS.
> And the popular consensus of the time was that PC-DOS was
> only a stopgap until CP/M-86 became available; at which
> point all the developers would port their new code back over.
IBM had a part in this. They asked the developer community
to "code to the hardware". This prevented developers from
moving back to CP/M 86.
In an interview with Gary Killdall in Kilobaud magazine (good luck
finding this one) he decribed his encounters with IBM. Appearantly,
they didn't want a multitasking multiuser operating system that would
harm the System 38, System 36, and the Series 1 market.
They wanted Killdall to sign nondisclosure and noncompete agreements,
effectively taking him out of the market he created.
This was in the "pre Gerstner" days, when there was so much corporate
inbreeding that the IBM executives thought the PC would be a good
"Terminal" for the Mainframe.
> >Essentially, there were several thousand "unknown soldiers"
> >who made the desktop PC a reality, and Microsoft did it's
> >best do destroy them all - decimating their revenues through
> >cutthroat retail pricing, and decimating them again through
> >tie-in per-processor licenses.
>
> While MS may seem to have low-balled the price of DOS 1.0 -
Actually, there seems to have been a bit of confusion about
the factors involved in the choice. Gates wanted to function
as an independent provider, offered nonexclusive rights for
very modest fees (not substantially less than CP/M), and IBM
may have thought they were getting AT&T's UNIX (Xenix). It was
a matter of public record that Microsoft owned Xenix and was
selling it to Tandy/Radio Shack on similar terms.
> (which is another factor in the cheapness of the PC)
Microsoft's biggest contribution to the cheapness of the PC was
that they got IBM to foot the bill for the development, promotional,
and distribution costs, then turned around and sold it to Compaq,
Kaypro, and a bunch of other Clone-Makers.
> they also had little in the way of incidental cost for that
> first release. Microsoft made its business selling the tools that
> promoted the growth of the PC platforms.
But those huge royalty checks from IBM didn't hurt either.
> And they were, at
> that time, far more interested in expanding the market for them
> to sell BASIC, Fortran and COBOL packages in, than in
> overriding CP/M as the most popular OS.
Very true. Microsoft was determined to put the other BASIC makers out
of business. One way they did this was to put their BASIC in the PC
ROM, the other thing they did was create a BASIC Compiler (BASCOM).
The real money maker was when they came out with their Version 2
product and told all of the ISVs that although Version 1 didn't
have a run-time license, Version 2 would have a $400 per USER
run-time license. I had been working for a company that had spent
about $2 million developing a legal time-keeping, billing, and
accounting package based on a shared network implementation. The
$400 run-time was a cost we had to eat, which was a problem since
we'd priced the product at $350 per user based on the assumption that
there would be no run-time costs. What made life worse is that
Microsoft refused to provide ANY further support for Version 1.
The company went Chapter 11, cut 80% of the staff, and was able
to stay afloat long enough to be a takover candidate to one of their
competitors. The product was ported to the PC, converted to C,
and sold as part of a larger package that included an office suite,
access to Lexus/Nexus, and worked as a loss-leader to generate business
for the "real" business. Eventually, as they began porting the product
to C, Microsoft reduced the run-time license to about $40/copy, but
only after a competitive substitute had been found.
> I don't think
> they were really aware of how popular that PC would become.
Actually, Bill Gates, Steve Jobs, and Gary Killdal were probably
the three people who knew how popular the PC would become back
when the MITS came out with it's row of 18 toggle switches.
> They had also learned well the business practices of the more
> successful companies such as MITS, Apple, Tandy,
> Commodore, and in particular, IBM and its competitors
> in the Mainframe marketspace. .
More importantly, Bill Gates had learned how to make the OEM
helpless without his software, and would then threaten to support
competitors unless the OEM paid Bill's price. It's likely that
IBM called Gate's bluff, and that's the real reason Compaq got
hold of MS-DOS.
Microsoft was still able to get nearly $10 per machine sold from IBM
and nearly $40/machine from everybody else. In 1991, IBM was the only
company to reject Microsoft's bundleware offer. That was because IBM
wanted to honor promises it had made to Lotus, WordPerfect, Corel,
and Borland to include their software with OS/2. Eventually, IBM
bought Lotus, Corel bought WordPerfect, and Borland merged with
Visigenics to become Inprise.
During this period, I was having to make key decisions affecting
about 50,000 users. I was acutely aware of what had happened,
how, and why.
> [ I think this is the point where we leave the refresher course ]
>
> >>Now the PC is more expensive, again thanks to Microsoft!
> >
> >Actually the same forces that created the economy of scale
> >that made Microsoft enabled PCs cheap have been moving from
> >Microsoft to Linux and UNIX.
>
> I'd argue that those are not exactly the same forces, and it is
> definitely not the same economy.
The forces I was referring to which you snipped, was the hobbiest,
the small business software developer, the independent programmer,
the staff programmer, and the consulting firms.
A big driving factor in favor of UNIX was that NT Server projects were
going into massive overruns. The average NT 3.51 project was canceled
before it was finished. The from 1997 to 1998 the average NT 4.0
project was 500% late and 800% over budget. Eventually, project
managers just factored in the extra into their estimates.
I remember talking with the VP of an IT department in a Fortune 50
company around 1995. He wanted to find a contractor who would
accept a fixed-price contract on an NT Server solution based on
estimates given for development on a UNIX server. I told him
that any contractor stupid enough to take the bid didn't have
the competence to do the job. Eventually, he found a contractor,
got his fixed price, and the contractor nearly went bankrupt trying to
meet the fixed price, and the product delivered was so poor that the
project had to be ported to UNIX anyway.
Most corporate IT departments have outsourced all of their development,
and most developers aren't willing to gamble on the instabilility of
Microsoft platforms. Even if the machine worked perfectly, Microsoft
would change the APIs and toolkits in the middle of the project.
In 1998, I spent several hours consoling someone who was leaving his
job as a project manager because he was so frustrated with the project.
He had started the project using COM, was then told he had to move it
to DCOM, and then he was told to migrate to ActiveX. He had to support
both Netscape Javascript and Microsoft Javascript, and then they
wanted him to deliver VBScript. The ActiveX VB Forms looked pretty but
didn't enforce the business rules, and when they tried to stress test
the system, it failed at 100 simulated users. Dll conflicts and the
removal of spinlocks during the installation of IE 4 caused terrible
crashes.
About a month later, his boss ordered the entire project ported to UNIX
as quickly as possible. Eventually, even the GUI interface was
converted to java servelets.
> Further, I think that the
> popularity of both Windows and Linux has had more of
> an effect on the UNIX-space, forcing many of the *NIX players
> out of the ivory-tower stagnation they had been feeding to
> the market for years.
Actually, it's even worse than that. The Ivory-tower guys lost control
of UNIX in the late 1980s. The IT shops that were using UNIX needed
fast, reliable, and flexible systems that could be maintained with
a skeleton staff. Ironically, one of the difficulties UNIX has in the
IT department is that the Windows administrators outnumber the UNIX
administrators by about 10:1 for roughly the same workload. This
means that there are dozens of Windows Server admins who have a vested
interest in the continued use of Windows. The UNIX team is usually
rediculously small. At one shop supporting 10,000 users for roughly
the same level of load, there were 250 NT servers, 40 NT
administrators, 25 UNIX servers and 5 UNIX administrators. The
UNIX servers were bigger boxes and maintences was managed primarily
through cron jobs, watchdogs, and shell scripts. The UNIX
administrators rarely left their desks. The NT administrators
were jumping all over the shop trying to find failed machines and
recover them before the users notices.
> >Microsoft stopped supporting the hobbiest programmer, creating
> >development tools that were so expensive that even the schools
> >couldn't afford them.
> Remember, the MSoriginal business plan
> was to *sell* programming tools...
Yes, originally, Microsoft wanted $500 for a 2 kbyte BASIC on paper
tape. The competition was charging $50 for a full function 8 kbyte
BASIC that did everything but MAT functions. The $75 Tarbell version
even did the MAT functions.
> (and in that, they found new competition
> with Phillipe Kahn's Borland
> low-balling the price of Pascal tools)
Microsoft has been plagued by horrible competition for decades.
Tarbell sold their BASIC for $75 when Microsoft was asking $500.
Lattuce sold their C for $50 when Microsoft wanted $200.
Borland wanted $50 for their Pascal when Microsoft wanted $300
for their Fortran.
PubSQL was $25 shareware when Microsfot was charging $200 for FoxBase.
And Linux was going for $50/copy for a fully functional multitasking,
multiuser operating system complete with Graphical user interface,
development tools, and server capabilities -----
when Microsoft was selling Windows NT 3.51 with notepad, calc,
paint, and quick-basic for $500. If you wanted the Server, the
price jumped to $1500 and you got IIS.
And NOW Microsoft want the Appelate court to believe that Microsoft
was responsible for bringing down the price of computers.
> >The kids weren't able to learn how the machine worked.
> >Those who stuck with the glamour of Microsoft
> >became "brilliantly ignorant", knowing the APIs, which changed
> >every 2-3 years, while knowing nothing about the protocols,
> >file formats, or the core workings of the systems.
>
> Add to that previous business plan a need to both make the
> interface easier for the end-users,
This Microsoft did well! In fact they did ALMOST as well
as APPLE did with the MACINTOSH. Of course I remember the days
when I'd have to teach people how to do a double-click.
Gates was smart. He insisted that users be able to use BOTH the
mouse AND the keyboard to perform the same functions. He insisted
that each button or icon have help available with the press of the
<F1> key.
> and make the machine more manageable for developers
This is where he blew it. DLL hell, shared libraries that change
after software is installed, lack of backward compatibility, and
addition of code that deliberately breaks code of competitors or
created using a competitor's product.
It's a developer's nightmere.
It keeps administrators up at night.
> - without the overhead of writing dedicated drivers
> for each application.
Kind of like the way Apple put all the documents out in Postscript?
Ever heard of DVI? UNIX had it almost 4 years before Microsoft
"invented it".
> Do you recall WordStar 6?
Gawd, we are going back a ways. Heck, I remember WordStar 2.0.
It was a bit like editing in troff/nroff.
> >Meanwhile Linux made it possible for next
> > generation developers to know
> >as much as they wanted to know about the machine.
>
> Much as DOS did in its heyday,
Absolutely. Third party developers liked DOS because they had
a pretty good idea of how it worked. Guys like Peter Norton reverse
engineered MS-DOS, and documented all the trade secrets Microsoft
didn't want us to know.
> as developers learned to work both with
> and around the OS to get things done.
Ahh, for the bad old days. Remember TSR's, software timing loops used
for copy protection, applications that jumped into the BIOS BASIC in
ROM which disappeared when the XT came out? There were some folks
with $200 software packages who were expecting free upgrades and were
quite upset when they had to shell out another $200 for the MS-DOS 2.2
compatible (XT) version.
> In both cases there is/was a need to differentiate
> the kernel hackers from the interface designers.
Back in those days, you had to be a kernel hacker to be
an interface designer. You could make BIOS calls to have
characters positioned on the screen, and you could even get
16 attributes. But if you wanted a real Lotus 1-2-3 interface,
you had to jump into the ROM.
Microsoft didn't really discourage this until just before it was
getting ready to release Windows 286. They assumed that the
developers would just start writing "well-behaved" programs and
wait until Windows was fast enough, stable enough, and ran on cheap
enough hardware to bring in the market.
> >The high-school hobbiests of 1993-1995, who cut their teeth
> >on Linux are just graduating from college this year.
>
> And are discovering that much of what they took for granted has
> started to adopt some of the same API ideas that DOS/Windows
> had to take in.
Linux has had consistent APIs all along. You have new and better
toolkits, especially with KDE and GNOME, but the core system has
remained pretty consistent.
> >These were the kids who created their own web sites using
> >80486/33 processors discarded by their fathers. Today, they are
> >designing clusters and control centers.
>
> Many of them are, and in some cases we are all the better for it.
> Others, unfortunately, are more interested in taking up a "cause"
> to completely change the digital landscape in favor of some
> new vision (but these, while vocal, are a minority).
Actually, the number in the first catagory is probably greater
than the number in the latter catagory. Most of the Windows-only
developers have been relegated to working as consultants on
limited development projects. Most companies that recently
bet their futures on Windows-only solutionss aren't doing too well.
Most of the Linux/UNIX developers have rapidly made the transition from
hobbiest Linux programmers to professional UNIX programmers. They have
been in high demand for E-Commerce, E-Business, B2B, and general Web
development. 75% of the servers on the internet are Linux/UNIX based,
and nearly all of the most profitable servers are based on Linux or
UNIX.
> >The companies Microsoft drove out of the Linux market weren't
>
^^^^^^
> I think you meant something else here?
> Maybe like "the CP/M, DOS/Windows market"?
Yup (shouldn't post at 2:00 A.M. :-).
> >completely dead. Many were merged to other bigger companies,
> >and many began supporting UNIX as early as 1991, and some began
> >supporting Linux as early as 1994. Rather than trying to
> >compete toe-to-toe against Microsoft for the desktop, they
> >developed applications that could be fronted by Web Browsers,
> >Java Applets, and trivial VB forms that put the real
> >application on the UNIX server. Putting the client and the
> >server on the same machine is a trivial matter with Linux or
> >UNIX workstations.
>
> While some companies have re-emerged
> in the client-server space,
> one can also say that competition
> there is fronted by companies such
> as Oracle and Sun, whose intentions
> are not clearly altruistic either.
This is true to some point, but then again, even this is a
wide open competitive market. Oracle has to compete with Sybase,
DB/2, Informix, Progres, and MySQL. Sun has to compete with HP,
IBM, Linux, and BSD for the server market, and has to compete with
PERL, Python, TCL, C++ and Yacc in the language market.
Java is nice, but contrary to Sun's argument, most large systems
can be built using stable tools and static servers - only the
5% of the code that runs 80% of the time neads to be optimized.
adding that code as a PERL or Python function is trivial in the open
source environment. The really intense part of the platform is
optimized for every application.
> >The OEMs have been tightly squeezed. When prices of PCs
> >dropped from $2000 per machine to $700/machine (not including
> >monitor) Microsoft actually tried to RAISE it's prices.
> While the range of prices for low
> to high end machines has come down, to
> claim that the average price for all
> PCs is now around $700 is a bit
> misleading.
The key is that the price for the core unit is the part
that really counts. This is where you compare apples-to-apples.
> I will not defend Microsoft's sales department, but I
> will point out that other items percieve as newer/better than
> previous versions (Monitors, CPU, Memory, etc) have been
> able to demand a higher premium even within the lowering
> of overall PC cost.
It's actually a bit funny. In 1991 I was asked to price a Sun
system and a Windows 3.1 system for a cost/benifit analysis.
The Sun IPC ran about $8000 with 8 meg of RAM, a 21 inch monitor,
and a 1 gig external SCSI hard drive which could be shared among
3-5 users. The Windows 3.1 machine cost about $6000 with 4 meg
of RAM, a 14 inch monitor, an 80 meg IDE hard drive, and a 4 MIPS
processor.
What we didn't know in 1991 was that the PC would have to be
replaced in 1993, 1995, 1997, and 1999 or 2000 to remain up to date.
Meanwhile, the Sun workstations just went from being the front box to
the back box. If you wanted a faster display machine, you could get
it, and simply use NFS and X11 to keep the benefits of your old box.
> >Today,
> >Linux users are willing to pay a premium for preinstalled Linux
> >and the cost of producing the machine is actually less than the
> >cost of producing a Microsoft machine.
>
> As long as you can provide the same level of peripheral support,
> and applications. Which will in turn raise the cost of production.
You may be right here. Linux machines are slightly more expensive,
probably because they aren't the "commodity" that Windows machines
have become. With Microsoft restricting the look-and-feel changes
to Windows 98, PC makers can only compete on price. With Linux
machines, they can compete on overall configuration.
The funny thing of course is that it costs the same to clone the
Windows drive as it costs to clone the Linux drive. Either or
both can be "printed" just as easily.
The big problem for most of the OEMs is that they are so focused
on looking good in the next quarterly report that they don't
get the chance to focus on the long term. Meanwhile, Bill Gates
was planning his "World Domination" in 1984.
> >Furthermore, Microsoft
> >turned the PC into a commodity, removing even the possibility
> >of product differentiation by forbidding boot sequence
> >alteration and desktop alteration on Windows 98.
>
> IBM turned the microcomputer into the commoditized PC. Most
> of the industry followed thier designs long before MS started to
> influence things. There are other companies that managed
> to differ themselves (Compaq, Dell, Apple), by either changing
> the business model, or not following the IBM-clone path.
Very true. From 1982 to 1989, IBM was holding up the train, and
Microsoft was extending their reach to other companies like Compaq
and NEC. There were some brilliant innovations such as the
Mac, the Atari 520ST and 1040ST, the Amiga, and UNIX boxes such
as the AT&T 6300 (which supported Tektronix 4010 graphics) the IBM
PC/RT, the Sun 1 and Sun 2, the Apollo, and the DecStation.
IBM was holding back the train until the release of MS-DOS 3.3,
at which time they were wanting real multitasking, GEM graphics,
and configuration independentn software more than anyone. That's
one of the reasons they wanted OS/2.
By the time X11R3 came out however, it was Microsoft who was
holding up the train. Microsoft had just sold it's rights to
the UNIX market as part of it's deal to sell Xenix to SCO. This
also included Linux (since it was a UNIX work-alike like Xenix).
X11 R3 came out about the same time as Windows 386. X11/R4 came out
about the same time as Windows 3.1 and created a really big problem
for Microsoft. In 1991, Microsoft was competing with $8,000
workstations, but it was pretty obvious that the $1000 UNIX box
wasn't far away. The death blow was when a guy from Finland
posted his operating system to the Free Software Foundation
FTP archive under the General public license.
By early 1992, the $1000 UNIX box (Linux) was a reality. While
Windows 3.1 and OS/2 2.0 required fast 80486 machines, Linux was
running quite comfortably on the 80386 machines being discarded
during the Windows 3.1 upgrades. By 1993, Linux had sprouted
full X11R6 support, a full suite of applications, and a
self-configuring interface that only required Video card tuning.
The one good thing about a Microsoft divestature is that
the Applications company could reenter the UNIX market
with things like MFC classes, and the Windows APIs.
They would have to compete in a wide-open market, but
the market itself could explode for everybody.
> >Meanwhile
> >Linux allows the selection from hundreds of permutations of
> >themes, color schemes, desktop managers, window managers, and
> >application preconfigurations.
>
> www.stardock.com, GEM, OS/2, BeOS, NeXT, GEOS, Desqview...
http://www.stardock.com/stardock/articles/oswars2000.html
I doubt Microsoft will allow stardock to live long. It would
be funny if the OEMs started going after it. Ohh that's RIGHT,
they can't monkey with the Boot sequence or the Display.
--
Rex Ballard - Open Source Advocate, Internet
I/T Architect, MIS Director
http://www.open4success.com
Linux - 90 million satisfied users worldwide
and growing at over 5%/month!
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
