[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Fwd: Article on Linux]




-- BEGIN included message

For the last two weeks I've been getting my toes wet with Linux. Thanks to a
few helpful Linux users on the net, including Glenn Shiffer, a member of
this mailing list (who is quite experienced with Linux, BTW), I'm up and
running. Two weeks ago I knew absolutely nothing about Linux. Here is a
brief history as well as a good comparison of the FREE Linux to the $800 NT
server. The "Linux vs. NT" debate looks a lot like the "NS vs. IE" debate.
Could MS be trying to swallow up Linux the way it has tried to swallow up
Netscape?
-------------------------------------------------------------------------------


Wired 5.08 By Glyn Moody

The Greatest OS That (N)ever Was

Linus Torvalds and thousands of disparate hackers created Linux,
it may be the greatest software story never completed.

Nor yet you seen
neither seen nor heard
when this earth was made
when the sky was built ...
-Kalevala 3:245-248


The Kalevala is an epic
poem of some 23,000
lines. Written in Finnish,
it forms a huge patchwork of
fragments taken from an oral
tradition evolving over thou-
sands of years in the province
of The Karelia, a region now split
between Finland and Russia.
Its creation in the middle of the
19th century was instrumental
in defining not only the Finnish
language but the Finnish nation.

The hero of the The Kalevala is
the shaman Vainamoinen. Early
in the poem he is challenged
by his hotheaded rival Jouka-
hainen to a duel of magic pow-
ers, expressed through secret
chants. At the end of this one-
sided contest, the upstart is
defeated, and the hero trium-
phantly asserts his unanswer-
able superiority by noting that
he, Vainamoinen, helped create
the world.

While the The Kalevala is the
result of an expansive literary
imagination, its origins resem-
ble the real-world inception
of what some consider to be
the greatest masterpiece that
the Internet has produced.
Aided by thousands of dis-
parate hackers  a "net.mind"
of sorts - Linus (pronounced
"LEE-nus") Torvalds, a 27-year-
old Finn, has also created a
new world. That realm, Linux
(pronounced "LINN-ooks"), is
an operating system, the force
that defines the digital domain
of programs just as our own
environment determines the
characteristics of animal and
plant life.

For most hackers, the goal is
to create neat routines, tight
chunks of code, or cool apps
that earn the respect of their
peers. Linus went much further,
laying down the foundation
that underlies the cool rou-
tines, code, and applications,
and achieving perhaps the
ultimate hack.

Linux was started six years
ago as a typical programming
lark: written to run on a pc with
4 Mbytes of RAM as a free ver-
sion of the costly commercial
Unix operating system. Today,
Linux has an installed base
conservatively estimated at
around 3 million users. And
they're not just spotty adoles-
cents playing in their bedrooms:
Linux vendors say that most
of the top companies in the
US have bought the OS - but
that few will readily admit to
running their multimillion-
dollar corporations on code
put together by a band of soft-
ware idealists.

Linux's installed base may
not be on the level of the 100
million-plus users of Windows,
or even the 50 million-plus in
the Apple Macintosh sector,
but Linux has made its mark
in just half the time. Linux is
freely distributable - one CD-
ROM can be passed on hundreds
of times - so it's particularly
popular in countries just get-
ting wired: South Africa, Cuba,
Mexico, Slovenia, Croatia, Rus-
sia, India, Pakistan, Nicaragua,
the Philippines, Bolivia. And
technologically, Linux eclipses
all the other brands of Unix.
"Linux is far and away the most
vital part of the Unix market,"
says Tim O'Reilly, founder of
tech book publisher O'Reilly &
Associates. Even Dennis Ritchie,
one of the two fathers of the
original Unix, calls Linux "com-
mendable."

The saga of Linux has many
strands. It is the story of Linus,
an arch-hacker of unusual wit
and charm who single-handedly
solved problems that usually
require teams of programmers
toiling for months. It is also the
story of the Internet as a model
for distributed collaboration.
Indeed, Linux is the Internet's
Kalevala, a huge patchwork
of code that defines a rapidly
growing cybernation, the tight-
ly linked community of those
who make and use it. What
unites these coders is the drive
to create the world's greatest
operating system, one more
powerful than any commercial
Unix, able to run on practically
any hardware, and infinitely
customizable. An OS, moreover,
that is fully the equal of Micro-
soft's flagship, Windows NT -
with true multitasking, virtual
memory, shared libraries,
TCP/IP networking, and other
advanced features. Many see
Linux as NT's most serious
competitor, the only viable
alternative to the Microsoft
monoculture - singular proof of
the ideal that says we should
have a digital choice.

But Linux also sits at a critical
juncture. Although the freely
distributable OS can never be
completely swallowed up by
Bill Gates's behemoth, Linux
needs to gain the trust of the
business sector if it wants to
take on the Microsoft machine.
Yet independence is a source
of fierce pride to Linux coders,
who fear Linux might become
just another honest freeware
Mosaic wiped out by a slick
commercial Netscape. In other
words, the same hacker ethic
that created Linux could cap its
expansion, making Linux the
greatest software story never
completed.

A kernel germinates

This epic begins in Helsinki.
Almost as if some cosmic hacker
had been at work on the source
code of life, it turns out that
Linus's home is just a 10-minute
walk from the center of the city
on a road called Kalevagatan
- Kalevala Street, more or less.

Kalevagatan offers a char-
acteristic mix of 19th-century
houses and blocks of modern
fIats. A tram periodically rum-
bles along the last part of the
street on its way down to the
Seaside Hotel, beyond which
lies the gray sea that surrounds
the city on three sides. Linus
lives with his wife in a building
that feels as if it were built to
accommodate college students.
Rows of bikes stand padlocked
under the stairs.

Linus himself looks more
like a schoolboy than a shaman:
medium height, with light
brown hair and blue eyes that
gaze out steadily from behind
glasses with roundish lenses
and mottled rims. Only the
eyebrows, which are remark-
ably dark and bushy,jar slightly
with the general effect of an
intensely boyish face.

Plenty of books line the walls
of his apartment. The place is
dotted with paintings and trin-
kets, rather chintzy curtains,
a desiccated crocodile between
a pair of armchairs, plus two
supercilious cats and a few com-
puters - three PCs, a Power Mac,
and three Alpha-based micros
lent by Digital - that lurk dis-
creetly at the edges of the room.
One of the most interesting
elements is barely noticeable:
a lead connecting the computers
to a telephone socket. This is a
256K leased line to the Internet,
installed and paid for by a local
Internet service provider as a
small token of gratitude to the
shaman of Linux.

Linux, it turns out, was no
intentional masterstroke, but
an incremental process, a com-
bination of experiments, ideas,
and tiny scraps of code that
gradually coalesced into an
organic whole. Many of Linus's
formative years of low-level
programming were spent por-
ing over a Sinclair QL, an eccen-
tric British computer launched
in 1984 that had many faults
but one real virtue: it was a
true multitasking system that
allowed advanced hacking. But
the key event that ultimately
led to Linux occurred in the
autumn of 1990, when Linus
took a Unix course at the Uni-
versity of Helsinki, where he
studied and eventually earned
a master's degree in computer
science this past February.

That fall the university had just
installed a MicroVAX running Ultrix.
Ultrix is one of the many flavors of Unix,
an operating system pervasive in university
science and engineering departments, and
routinely used by corporations for heavy-
duty computing. Unix is also inextricably
linked with the history of the Internet
in fact, it still runs most of it. created in
AT&T's Bell Labs in 1969, Unix grew in the
1970s as a labor of love, the product of
devoted hackers such as Ken Thompson,
Dennis Ritchie, and Bill Joy. In 1993, AT&T
sold Unix to Novell, which sold it to the
Santa Cruz Operation in 1995. Today, vari-
ants are available from companies such
as SCO, IBM, Digital, HP, and Sun, resulting
in a highly fragmented market - one that
is increasingly vulnerable to Microsoft's
rival system, Windows NT.

But in 1990, Linus was occupied with
more mundane matters: his university's
hardware couldn't cope with more than
16 users at a time."You had to wait in line
to get to a terminal," he recalls.

One of his course books was Andrew
Tanenbaum's Operating Systems: Design and
Implementation, which provided a guide
to a kind of baby Unix called Minix."That's
when I actually broke down and got a PC,"
Linus says. Until then he had resisted, he
explains, because "if I had gotten a PC, I'd
have gotten this crummy architecture with
his crummy MS-DOS operating system and
wouldn't have learned a thing."

Minix was very limited, designed simply
to teach operating systems, explains Tanen-
baum, currently professor of computer sci-
ence at Vrije Universiteit in Amsterdam.
Nonetheless, he says, Minix seemed to strike
chord:"Within two months of its release
in 1987, there was a newsgroup with over
40,000 users worldwide."

However, many wanted more capabilities.
I was getting hundreds of emails a day
asking to add this feature and that feature,"
Tanenbaum says."Many people were frus-
trated with my constantly saying no."

Linus began to experiment with his own
hacks, using Minix as scaffolding to develop
a new program."I made two processes and
made them write to the screen and had
a timer that switched tasks," he recalls."One
process wrote A, the other wrote B, so I saw
AAAA, BBBB, and so on."

Linus says he never intended to create
a kernel, the part of an operating system
where the real processing and control work
is done. Instead, a purely practical need
to read Usenet newsgroups drove him to
modify those first two trivial processes.
"At some point," he recalls,"I just noticed,
Hey, I almost have this functionality."

Linus had been a true hacker early on:
in his early teens he had programmed a
Commodore Vic-20 micro the hard way,
using assembly language - partly because
he didn't recognize there were other tools
available, and partly because it just seemed
natural. In 1991 he needed a simple terminal
emulation program to access newsgroups.
So Linus sat down and wrote one  based
on his two-process lash-up. As Linus tells
it, doing so was simply a matter of changing
those As and Bs into something else."One
process is reading from the keyboard and
sending to the modem"  which then con-
nects to the university computer  "and
the other is reading from the modem"
- receiving the newsfeed -"and sending
to the screen."

But there was something else Linus
needed: drivers. A driver acts as a software
buffer between something central (like
the kernel) and something peripheral (like
a keyboard, screen, or modem). You could
build this straight into the kernel, but then
you'd have to rewrite the kernel each time
you wanted to use a bigger screen, a differ-
ent keyboard, or a faster modem. Far better
to abstract out this layer in a driver: to use
a new screen or keyboard, you just slot in
the new driver and the old kernel works.

In the summer of 1991 - just six months
after he got his first pc - Linus found he
needed to download some files. But before
he could read and write to a disk, he recalls,
"I had to write a disk driver. Then I had to
write a file system so I could read the Minix
file system in order to be able to write files
and read files to upload them," he explains,
as if it was the only reasonable thing to
do."When you have task-switching, a file
system, and device drivers, that's Unix"
- or at least its kernel. Linux was born.

This fledgling system would have been
short-lived had Linus not mentioned it
in the Minix newsgroup. His early posting
prompted an offer of space on an FTP server
at the Helsinki University of Technology,
letting people download the first public
version of Linux."Linux was my working
name," Linus says,"but if I actually used it
as the official one, people would think that
I was an egomaniac and wouldn't take it
seriously. So I chose this very bad name:
Freax" - free + freak + x."Sick, I know." Ari
Lemmke, who ran the FTP site, decided he
didn't like the Freax label, so he used the
working name instead.

By January 1992, only 100 people or so
were using Linux, but they provided a criti-
cal online baptism. Those early uploads
and comments were crucial. Particularly key
were the patches sent in by fellow hackers
to fix problems they found with the code.
By chance Linus had stumbled into an
online Karelia and was about to start piec-
ing together the patchwork of his Kalevala.
Anybody anywhere on the Net could obtain
the basic Linux files. Email enabled them
to comment and offer improvements, while
Usenet provided a forum for discussion.
Beginning as the product of one mind,
Linux was turning into a tapestry, a move-
ment of like-minded hackers.

The sorcerer's apprentices

A kernel on its own is not much use, even
if it is being refined constantly through
patches sent by interested hackers. Part of
the reason Linux took off so spectacularly
is that nearly everything else needed for
a complete OS was there waiting.

These programs-in-waiting were part of
the Free Software Foundation's GNU proj-
ect. The recursively named effort -"What's
GNU? GNU's Not Unix" - was begun in 1984
by Richard Stallman as a reaction against
some of the draconian rules imposed on
software users by vendors. (See "Is Stallman
Stalled?" Wired 1.1, page 34.)

GNU's aim was to write a complete"free"
version of Unix - the kernel and all the
associated elements - that is, one that
gives users the freedom to share and
change software but not add restrictions
and impose them on others. With the Linux
kernel, Stallman says,"the available free
software added up to a complete system."

Rather than wait for someone to write
applications designed specifically for his
operating system, Linus tweaked Linux
to perfectly fit GNU's pre-existing apps.
"I never ported programs," Linus says.
"I ported the kernel to work with the pro-
grams. Linux was never the primary reason
for anything - user programs have always
been the reason."

A similarly pragmatic approach allowed
Linux to acquire, almost overnight, a graph-
ical front end a la Windows - this was indis-
pensable for its wider acceptance. (Until
then, Linux was controlled through obscure
commands entered as text at a prompt,
rather like DOS.) The GUI was provided by
the Xfree86 Project, a nonprofit group that
provides free software for pc versions of
the X Window System.

Linus also adopted the standard GNU
licensing scheme called copyleft. The gen-
eral public license, or GPL, allows users to
sell, copy, and change copylefted programs
- which can also be copyrighted - but you
must pass along the same freedom to sell
or copy your modifications and change
them further. You must also make the
source code of your modifications freely
available.

GPL has proved a powerful force for
Linux's success. First, it has encouraged
a flourishing commercial Linux sector.
Although Linux is readily available on the
Internet, buying a set of CD-ROMs for US$30
is generally far cheaper than downloading
several hundred Mbytes of code - and far
quicker. GPL also has given programmers
an additional incentive to join in the
essentially philanthropic spirit of the
Linux movement. The license has ensured
that their work would be freely distrib-
utable, but not unfairly exploited or locked
into proprietary products by unscrupulous
commercial organizations.

In a sense, GPL provided a written con-
stitution for the new online tribe of Linux
hackers. The license said it was OK to build
on, or incorporate wholesale, other people's
code - just as Linux did - and even to make
money doing so (hackers have to eat, after
all}. But you couldn't transgress the hacker's
fundamental law of software: source code
must be freely available for further hacking.

In March 1994, the official Linux 1.0
appeared, almost as a formal declaration
of independence. By then the user base was
already large, and the core Linux develop-
ment team substantial. Among the thou-
sands of files Linux contains, there is one
called simply Credits. In it are the names,
addresses, and contributions of the main
Linux hackers. The list runs to more than
100 names, scattered around the world.
Almost uniquely for a hacker project, Linux
has huge and comprehensive sets of FAQs,
how-tos, and general help files (see, for
-example, the Linux Documentation Project).

The growth of the development team
mirrored the organic, not to say chaotic,
development of Linux itself. Linus began
choosing and relying on what early Linux
hacker Michael K. Johnson calls"a few
trusted lieutenants, from whom he will
take larger patches and trust those patches.
The lieuts more or less own relatively large
pieces of the kernel."

The Linux approach is deceptively simple.
All hackers are free to work on any addi-
tional features or improvements. Even in
the earliest stages, new code is freely down-
loadable for users to try and critique: beta-
testing is not a last-minute attempt to catch
the worst flaws, but an integral part of the
process. When several people work on the
same area, they may compete or combine;
if they compete, the best code wins through
raw Darwinian selection.

"The Linux and free software community
can be thought of as true meritocracy,"
says Marc Ewing, who in 1994 founded
Red Hat Software, which sells one of the
most popular Linux distributions."People
in a traditional development group are
assigned jobs that they may not know
much about, or be best suited for."

Bruno Haible, who has contributed to
Linux's memory management code, puts
it even more succinctly:"When the main
author doesn't improve his code anymore,
other people will."

This freewheeling situation has allowed
hundreds of thousands of users to employ
Linux on perhaps tens of thousands of
hardware configurations: Linux supports
everything from an lntel 386 to a Pentium
Pro, along with platforms based on Alpha
(Digital's RISC chip), SPARC (Sun's RISC chip),
MIPS (port to Silicon Graphics's RISC chip
under development), and MkLinux (a ver-
sion of Linux that runs on Intel and Power-
PC machines).

Users also have, via the Net, ready means
for communicating any problems to the
person who knows the program best - the
author. That can be a big plus, or a distinct
Linus. While serious hackers might like
having a tete-a-tete with another code-
jockey, most regular users just want their
questions answered fast. Providing a reli-
able Linux help desk could help a commer-
cial vendor bring Linux to the masses.

Linux, of course, is not alone in using the
Internet for distributed development and
user feedback.:"The main difference,"says
Eric Youngdale, a coder who has led the
development team for Linux's SCSI drivers,
"was basically Linus."Free-flowing self-
regulation is all very well, but without the
right person to act as a focus, this energy 
will just be dissipated. Linus is omnipres-
ent through the development approach
he created. Yet he almost never intervenes
- in a way, he solved all the problems up
front.

The automatic selection of programmers
to work in the areas they know best, and
the ability of the system to expand end-
lessly by delegating tasks in this naturally
distributive way, has produced other bene-
fits."I am impressed by the speed at which
Linux has obtained features that have taken
commercial vendors many years to
develop," says jon "Maddog" Hall,
senior marketing manager for the Digital
Unix Software Group and executive direc-
tor of the nonprofit Linux International.
Indeed, the pace of upgrades is vertigi-
nous: from the earliest days, the latest
patches have typically appeared every
week. And yet in parallel, there is always
a stable release distribution that moves
forward just as inexorably when the new
features have been thoroughly tested. Linux
generally proceeds with point releases
- 1.1 , 1.2, et cetera. There is also a compli-
cated system of subpoint releases such
as 1.1.12. When a big enough jump in soft-
ware functionality occurs, developers move
to the next version, a process normally
presided over by Linus.

This two-track development process
has made Linux probably more advanced
and yet more stable than any other version
of Unixes today."Linux is now entering an
era of pure development instead of just
catching up," says jacques Gelinas."This
is where the huge number of developers
and testers will give an incredible edge
to Linux over any other operating system
- NT included,"

Phil Holden, Microsoft's product manager
for Windows, seems not to be unduly con-
cerned: "We have a very talented team of
developers making sure NT is the most
powerful, flexible, and easy-to-use operat-
ing system," he says.

But staying on top may prove increasingly
difficult for Microsoft. The latest version
of Linux - release 2.0 - offers 64-bit process-
ing (NT and many Unfixes are only 32-bit);
symmetric multiprocessing, which allows
the simultaneous deployment of several
chips in a system; and networking more
advanced than that of any other operating
system.

A related advantage of Linux's develop-
mental structure is that security fixes typi-
cally turn up faster than from commercial
suppliers. For example, when a "Ping of
Death" assault of multiple, low-level mes-
sages crashed several operating systems
worldwide, a quick patch to Linux enabled
the attack to be thwarted in a couple of
hours."Somebody posted a report of the
ping," recalls Alan Cox, author of the fix,
"so I just sat down, fixed it, and posted the
fix straight back." Users of other operating
systems had to sweat out their vulnera-
bility far longer.

Covert code

While the quality and sheer technical bril-
liance of Linux are undisputed, few people
outside its circle have heard of it. Linux
could easily seem to be some interesting
but ultimately marginal phenomenon.
The facts suggest otherwise,

Because Linux can be downloaded from
hundreds of sites and users are encouraged
to pass on CD-ROMs, it is impossible to accu-
rately determine how many people use the
OS. Based on an amalgamation of voluntary
registration systems and market research,
however, Linux distributor Red Hat esti-
mates that there are between 3 million and
5 million users worldwide. These numbers
understate the true reach of Linux, accord-
ing to Harald Alvestrand, project coordina-
tor of The Linux counter."The more than
46,000 users who have chosen to register
are less than 5 percent of the total number
of Linux users," he estimates, extrapolating
from a comparison of the Counter's figures
and the considerably higher number of
Linux machines believed to be running Web
servers."I figure the true ratio is closer to
0.5 percent. That would indicate 9 millpond
Linux users."

Whatever the actual numbers, it is not
unrealistic that Linux will emerge as the
second operating system after Windows,
especially given Apple's currently confus-
ing sense of direction. However, Micro-
soft's huge installed base of users and
wide range of applications is unlikely to
be surpassed, despite what believers in
the Linux magic might hope. Linus has no
illusions on the subject and even employs
tools like Microsoft's PowerPoint - running
on top of Linux via one of the Windows
emulators available.

Yet Linux's importance lies not just in
the size of its installed base, but also where
those users are found."More than 1 20
countries are represented," according to
Alvestrand."And Linux is a real power in
the less developed countries - in some
cases growing faster than the Internet."

Given the growing expanse of users
working collaboratively, today's Linux is
less a seamless piece of coding than a
tapestry of hundreds of hackers' contri-
butions. The parallels with the Kalevala
are striking. Finland's national epic was
created in the middle of the last century
when a district medical officer, Elias Monn-
rot, traveled around Karelia collecting lines
of poetry from many different sources -
scraps of an ancient oral tradition - and
wove them together. When Linux first
emerged, it consisted of some 10,000 lines
of code. Now the kernel alone is almost
1 million lines, with millions more in the
hundreds of ancillary programs that make
up a full Linux distribution.

Given that Linux is immensely powerful,
free, and truly open, you might expect cost-
conscious companies to be rushing to
embrace it, to obtain more software for little
money and without the crippling depen-
dence on one supplier that the Microsoft
approach implies. And they are - in secret.
Freeware is a dirty word in most companies.
And certainly not something on which you'd
base mission-critical operations.

Bryan Sparks heads Caldera, a company
that hopes to claim its piece of the business
market with OpenLinux. Caldera was set
up with funding from Ray Noorda, the for-
mer CEO of Novell and a longtime critic of
Microsoft. There is no love lost between
the two companies - Caldera has filed an
antitrust lawsuit against Microsoft relating
to DOS coding allegedly pinched from
Caldera-owned DR-DOS. Sparks definitely
sees Windows NT as a competitor."You
have to if you're selling an OS - Microsoft
is the 2,000-pound gorilla," he says. But,
Sparks notes, Caldera focuses on small- and
medium-sized businesses, typically those
with fewer than 50 servers."We don't pri-
marily target companies like Ford
Motor Co., although we've sold some
there," he says."We look at ourselves as a
grassroots company."

OpenLinux costs anywhere from $60
to $400, but Sparks still must frequently
calm user fears about a product primarily
regarded as a hacker group project."We
had an MIS guy at a large company who
called and said,'We're using Linux for gate-
ways and other things to attach all our
Internet sites. My MIS director just found
out that I'm running Linux and said we're
not going to run our company on no free
software. Can you help me?' And I said,'You
bet we can.' Linux is free, but we're a real
vendor; if you have problems we'll stand
up and take the bullet."

Red Hat's Marc Ewing says his company
already has many top-drawer customers,
including NASA, Disney, Lockheed Martin,
Industrial Light & Magic, General Electric,
Ernst & Young, Price Waterhouse, UPS, Nas-
daq, the IRS, and Boeing, as well as leading
US universities.

Skeptics contend that Linux will never
break into the mainstream because of
its lack of desktop applications. Even as
recently as a year ago, this was probably
true. Linux hackers - including Linus - gen-
erally don't seem interested in writing apps;
Mnemonic, a Linux browser project, is a
recent exception. But now that the Linux
market is expanding, commercial compa-
nies are filling the gap. Applix and Star
Division both offer integrated suites of
powerful desktop programs - wordproces-
sors' spreadsheets, graphics applications,
et cetera. Caldera has put together a Solu-
tions CD, including products by big names
such as Corel (a version of CorelDraw to
add to the existing Linux port of Word-
Perfect) and German company Software
AG (Adabas D database).

Other major firms that have ported
or helped others port to Linux include
Netscape (Navigator 3.0 browser and Fast-
Track Web server) and Adobe (Acrobat
PDF reader), plus companies like Marimba
(Castanet). Even Microsoft is coming out
with a Linux version of its standard for dis-
tributed computing, DCOM (Distributed
Component Object Model, also known as
network OLE - basically a way of linking
distributed software objects across net-
works). And Linux's java support means
that any apps produced in this platform-
independent language will immediately
add to the pool of Linux software.

However, it is worth considering some of
the obstacles. One is that the distributed
development model may ultimately fail,
"As time goes on," Eric Youngdale confesses,
"it gets harder for key developers to com-
municate because of the enormous noise
on the mailing lists." This could provide the
opening needed for a major developer to
pull Linux more completely into the com-
mercial sector.

Even if the development process contin-
ues to feed new improvements into Linux,
this in itself may be ultimately destructive.
One of Linux's great advantages has been
its leanness, but the temptation to add more
and more features can end up causing soft-
ware bloat - huge programs trying to do
too many marginal things.

And then there is Microsoft. With Bill
Gates steadily buying up most of the OS
talent around, can Linux hope in the long
run to compete against Windows NT? One
reason it might is that Linux 1.0 came into
existence more or less fully formed - like
Vainamoinen's world - as a complete,
advanced system, rather than one that
evolved slowly and painfully through vari-
ous primitive stages. NT, on the other hand,
was designed to be the robust incarnation
of Windows 3.0, itself the third iteration of
a program that began as a hack to make
MS-DOS usable. MS-DOS, in turn, was heav-
ily based on a very simple operating system
that Gates bought from Seattle Computer
Products in 1981 just before IBM launched
its pc - and MS-DOS.

Linux, by contrast, was able to incorporate
all the past successes of Unix and ditch its
failures - another reason Linux may stand
a chance against NT."Given the long history
and millions of person-hours of unpaid
efforts that are crystallized in Unix,'. says
Dick Pountain, contributing editor to Byte
magazine,"the resources that lie behind
Linux may be even greater than what Micro-
soft has put into NT,"

However, Linux still needs to address
the crucial issue of how to gain the trust
and confidence of the business sector,
And Caldera's example could be key. The
question is, Can it successfully mediate
between the distributed Linux ethos and
the demands of the business world - and
can this approach be scaled to match the
Microsoft machine?

Linus 2,0

Suddenly, at the end of 1996, amid the
growing sense that Linux may well be
approaching a historic juncture in its pro-
gress, came a stunning announcement,
Linus was leaving the University of Helsinki
to work for an unknown US computer com-
pany called Transmeta, based in Santa Clara,
California, Many initially wondered whether
Linux would be fatally wounded, But senior
developers and commercial vendors now
seem fairly confident that Linux now has
enough momentum and structures in place
to continue.

What's Linus going to do at Transmeta?
"I can't tell you," he says coyly."But it's
actually in my contract that I'm doing Linux
part-time."

Though Linus nominally spent the last
seven years at the University of Helsinki
working on his master's degree, in reality
he devoted much of that time to Linux, He
and everyone else knew that one day his
thesis (on porting Linux) would be finished
- and with it his research salary. Then he
would confront the perennial question for
aging hackers: How do I earn a living in the
real world?

Some, like Matt Welsh, remain in academe;
others, like Hannu Savolainen, now sell
commercial versions of their Linux work
- in this case, drivers for sound cards, set-
ting up his own company would have been
one option, but Linus was not interested:
"At this point,"he says,"I wouldn't want
the paperwork." He happily concedes
a quite unhackerish desire"to have
money," but adds"it's not my primary goal
in life."

Basically, Linus needed a new challenge,
and he candidly admits that"if things at
Transmeta go really exceptionally well, I'm
going to be rich just working and doing
whatever I want anyway."The only techni-
cal detail he will divulge about Transmeta
is that it is into VMSI, or very large-scale
integration - chips, in other words, though
"they obviously have a software side, too."
Transmeta, it turns out, is a start-up
headed by Dave Ditzel, chief scientist of
the chip development project at Sun that
produced the SPARC processors, probably
the most successful example of the RISC
idea. One of Transmeta's big investors is
Paul Allen, the other founder of Microsoft.

Although there are no public statements
(www.transmeta.com/ offers only the self-
contradictory declaration,"This Web page
is not here yet"), Paul Allen's Web site men-
tions something about Transmeta being
involved in the creation of"alternative
VMSI engines for multimedia PCs."

But despite the fears of some in the Linux
community, it is not Transmeta that is likely
to distract Linus, but another, more per-
sonal, genesis. He has an infant daughter
- Patricia Miranda - born in December 1996,
two days before we met.

Ever the realist, Linus sees his daughter
- more than his move to Silicon Valley -
as likely to change his involvement in the
development process of Linux. And what
right have we to complain if the shaman
who conjured up the rich world of Linux
2.0, the ultimate digital hack, should now
want to concentrate on his other creation
- Linus 2.0 - the ultimate analog hack?

Besides, if this latter-day Vainamoinen
ever wishes to renounce completely his
leadership of the online tribe, now there's
sure to be someone out there ready to
lead the Linux nation into battle.



That is where the way goes now
where new track leads
for more versatile singers
more abundant bards
among the youngsters rising
among the people growing,
-kalevala 50:611-620

--------------------------------------------------------------------
Linux related sites

Linux Documentation Project             http://www.sunsite.unc.edu/LDP/
Linux International                     http://www.li.org
Linux Kernel Archives                   http://www.kernel.org
The Linux Counter                       http://www.munken.uninett.no:29659/
Free Software Foundation's GNU Project  http://www.gnu.ai.mit.edu/
Xfree86 Project                         http://www.xfree86.org
Transmeta homepage                      http://www.transmeta.com/

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

    Linux          Versus          NT
 
   Linux 2.0                 Windows NT Server 4.0

----------------------       ----------------------
System Requirements          System Requirements
Intel-Compatible System      Intel-Compatible System
------------------------     ------------------------

o 386/16-MHz,Pentium,or      486/33-MHz,Pentium,or 
  Pentium Pro Processor      Pentium Pro Processor

o 4 Mbytes RAM,              12 Mbytes of RAM
  16 M recommended           16 M recommended

o 20 Mbytes hard-disk in     125 Mbytes hard-disk
  character mode, 100 M      
  to install X Windows

o FTP, CD-ROM or access       CD-ROM
  to CD-ROM via Network

o                             VGA, Super VGA
                              Compatible vid graphics adapter

------------------------      ----------------------------
RISC-based system             RISC-based system
------------------------      ----------------------------

o 16 Mbytes RAM               16 Mbytes RAM
  32 M recommended

o 100 Mbytes hard-disk        160 Mbytes hard-disk
  in char mode, 180 M
  with X Windows

o FTP, CD, Network            CD-ROM

o                             VGA, SVGA, Vid Adapter

-------------------------     -----------------------------
FEATURE SET                   FEATURE SET
-------------------------     -----------------------------

o True 32-bit system on       True 32-bit system on Intel
  Intel and SPARC (64-bit     and Alpha
  port for UltraSPARC 
  underway), 64 bit on
  Alpha

o Supports 1 to 16 users      Multiuser
  Pentium/Pentium Pros as
  a multi-processor system

o Virtual memory              Virtual Memory

o True protected-mode         True protected-mode
  multitasking                multitasking

o Monolithic kernel           Microkernel architecture
  architecture

o Memory protection for       SMP support for up to 32
  apps and subsystems         processors per server

o Ethernet,PPP,SLIP,NFS       Memory protection for apps/subsys 

o Network firewalling         Modular hardware drivers
  Network aliasing
  (Virtual Hosting)

o                             Ethernet,PPP,SLIP

-----------------------       ----------------------------------
Networking Options            Networking Options
-----------------------       ----------------------------------

o SMB                         SMB
o Appletalk                   Appletalk
o Novell NetWare 3.x          Novell Netware
o NFS networks                NFS Networks
o TCP/IP networks             TCP/IP Networks
o                             Microsoft WFWG
o                             IBM SNA networks
o                             RAS by way of ISDN,X.25
                              and standard phone lines

-----------------------       -----------------------------------
Cost                          Cost
-----------------------       -----------------------------------

o Free. Commercial            $809 (five-client license)
  versions available
  from US$30; No limit
  on number of users

-------------------------------------------
Glyn Moody (glyn_moody@cix.compulink.co.uk) 
divides his time equally between London,
northern Italy, and cyberspace.

http://cybercom.net/~pleonard 
    
---------------------------------------------------------------
Boston Online Commands
To: majordomo@cybercom.net
Body: subscribe boston-online
Body: unsubscribe boston-online

-- END included message