UKUUG talk 9th August: Ted T'so Linux: Past, Present & Future

[Chris Benson]

Hmmm, it appears that certain words in the first paragraph of a message
are trapped by the mailing list software as misdirected admin commands.
I've just had 2 out of 3 of my messages redirected back to me for manual
correction. 

Perhaps this will allow the message to get though!

----- Forwarded message from Chris Benson <chrisb at jesmond.demon.co.uk> -----

Date: Sat, 11 Aug 2001 17:43:09 +0100
From: Chris Benson <chrisb at jesmond.demon.co.uk>
To: tyneside-pm at happyfunball.pm.org
Subject: UKUUG talk 9th August: Ted T'so Linux: Past, Present & Future

Tim O'Reilly was booked to give a talk "The pause before the next wave"
about what happens next to the high-tech. economy.  He had to cancel,
but a long-time Linux kernel hacker Theodore T'so stepped in to talk
about Linux: Past, Present and Future.

Some misc. notes.

Using the start-off point that Linux was about 10 years old (the exact
date depends whether you count birth or conception -- as Ted pointed out
this is possibly more important in the US than in Europe :-)  he gave
us some history.  Pointed out that a lot of things happened as long as
5 years ago ...

He talked about the Linux creation myth: write a process switcher, test
it: AAAABBBABABABBABAAABABBAAAABBBBBB, add a terminal driver and then
a few system calls later you have a Unix-like o/s.

He talked about the present: companies large and small faring well,
and not so well in the Internet/Linux age.

Business Models

Some do not work - banner ads, building/selling hardware to add value to
Linux (Ted works for VA/Linux that has just pulled out of hardware), ...

The hobbyist/enthusiast will always be around which ensures the survival
of Linux.

IBM get more revenue from services than from hardware, therefore Linux
was not a threat (and if Linux reduces NT sales by running on IBM's intel
hardware that's fine,  and if Linux reduces sales of Suns by running on
S/390 hardware, that's fine too).  Predicts that IBM will do well ...

Another model: BitKeeper source control software: written by Larry
McVoy in his own time.  Available to anyone who wants to use it *BUT*
the change files are made public unless you pay for it.  The software
is good and works and supports a 12 person company.

There was a digression into what BitKeeper does and how it works ... and
into how Linus is one of those people who is so naturally good they
haven't had to learn how to deal with software engineering problems
that mere mortals have to cope with (Linus does not use any source code
versioning s/ware on Linux,  though others do).  BitKeeper is the business
(if he should choose to use anything!).

The talk about the future became (of course) less specific.  A lot
of it around political, legal and social issues: copyright, patents,
trade secrets.  These may kill open source: if it becomes illegal to
reverse engineer s/ware to interoperate, away goes SAMBA and much else.
If more stupid patents are granted it gets more difficult for small
companies (and academics and enthusiasts) to write any code -- only
corporations with armies of lawyers *and* large sheaves of their own
patents to bargain with, will be able to write code without being sued.
Of course Open Source developers are at a disadvantage because they make
the code available ...

Linux (and other open source developers) need to produce code that people
can use (story about being unable to build a Debian package of the kernel
because there is a ' in his name ... and if someone whose been building
Linux kernels for over 9 years can't do it ...).  This may need usability
testing and regular re-testing, standards documentation, ....

While some types of open source progress very quickly, some areas (SCSI
subsystem, documentation, ...) don't.  Need to work out a way of getting
the grunt work done.  This may take social engineering -- not a common
hacker trait.

One estimate puts Linux on 5% of desktops: about the same as MacOS.
If Linux can provide a stable platform (not everyone wants to compile
and install a new libc to run some app),  it could be worth s/ware
companies supporting Linux.

Become politically active: tell politicians (in language they
understand) about the effects of their laws.

And that was about it. 
-- 
Chris Benson

----- End forwarded message -----

-- 
Chris Benson