See also: last week's Back page page.

Linux links of the week

A similar sort of site is ionline. They, too, concentrate on interesting articles and tutorials, in a more subject-oriented way. The current set is mostly about PHP3 and its uses.

Letters to the editor

Date: Thu, 01 Oct 1998 10:49:26 -0500
From: Craig Goodrich <craig@airnet.net>
To: editor@lwn.net
Subject: Red Hat's growth ...

In your 10/1 issue you comment

> All of these [developments connected with RedHat's growth and
> additional investment] can be seen as good things ..., but all
> together they raise the question of just how big and influential
> Red Hat should get... We certainly do not need a single company
> in a Microsoft-like role with an overwhelmingly large market
> share. 

While no reasonable tuxie could disagree that the diversity
and competition among distributions is one of the best things
about Linux, I think your concern here is exaggerated.

First, if one distribution is to become a favorite, a good
case can be made that none deserves that position more than
Red Hat.  They have consistently pushed the state-of-the-art
forward for the Linux community, from the development of 
RPM years ago through today's funding of such superhackers
as Alan Cox and the breaking down of chipmakers' NDAs for
programming information.

I recall several years ago telnetting anonymously into
Linus' machine in Helsinki and being greeted by a Red Hat
login banner (for RH2, I think).

Second, Red Hat has always kept closely to both the letter
and the spirit of the GPL -- and in the process allowed new
competitors to spring up based on Red Hat's own distribution.
SuSE, for example, began as a Red Hat distribution customized
for the German market.  The original Caldera was based on
Red Hat.

Any creator of a Linux distribution has innumerable decisions
to make, some major -- e.g. libc5 or glibc? -- and many minor
-- do I go with the latest release of xbunnies or use the more
heavily tested older version?  In the Linux world, if there is
widespread controversy over any of these decisions, it's likely
that someone will address that concern with yet another distribution:
For those worried about the stability of glibc, LinuxPro is 
available, a slightly tweaked Red Hat 4.2.  For those (like 
this writer) who believe the Red Hat position on KDE is just 
silly, the Mandrake version of Red Hat 5 recently appeared.

The free market is an incredibly powerful force, and when 
combined with the quirky ethos of the free software movement,
it assures us of continuing wide choice in Linux distributions.  
The only way Red Hat could gain and hold a monopoly position is
by putting out a distribution that made *everybody* happy.  And,
as the most casual perusal of Slashdot or c.o.l demonstrates,
that jes' ain't gonna happen.

So yes, indeed, best wishes to the other distributions.  Some
will die off and new ones will appear, of course, but let's not
worry the community about it.  Our Linux choices have multiplied
enormously during the very period that Red Hat was establishing
its premier position among the distributions, so as Linux is
finally beginning to receive its well-deserved attention in the 
business community, let's just relax and enjoy the ride.

Sincerely,

Craig Goodrich
Rural Village Systems
somewhere in the woods
  in Elkmont, Alabama
			===========
There are two ways of constructing a software design. One way is
to make it so simple that there are obviously no deficiencies
and the other is to make it so complicated that there are no
obvious deficiencies.
			--  C A R Hoare

Date: Fri, 2 Oct 1998 15:09:13 +0200 (MET DST)
From: Maurizio de Cecco <Maurizio.de.Cecco@ircam.fr>
To: editor@lwn.net
Subject: Applications, the Next Challenge


Can Microsoft (or anybody else) takeover Linux ?

We all know the answer: no.

But this "no" refer to the kernel; what if we consider the user market ?

In this case the answer is simply yes.

For example, take this strategy illustrated in a discussion in
SlashDot: Microsoft port MS Office to Linux, together with a Window
Manager or Desktop that provide proprietary undocumented extensions,
not friendly to non-MS applications.

Or, just consider the porting of MS Office, without any other
monopolistic strategy; will this change anything for the average 
user ?

No, everybody would still be caugth in the update-tax paradigm, the
monopole on the office suite market will just be renforced; and, if
you consider that in the last two years Linux got far more new users
than MacOS, and that MS-Office exists for MacOS, you will see that is
not at all unreasonable to have a MS-Office suite on Linux (expecially
if Linux really win the war against NT).

What the Linux community can do to prevent these mechanism to
replicate themselves on Linux ?

I don't know, but for sure it is time to seriously think about all
this.

The RMS position is one answer: develop free applications that are
competitive with the commercial one, and have the Open Source movement
to win the application war after the OS (ok, this is not the RMS position,
this is my personal interpretation of it :).

Personally i think we need on thing more: open formats; for example,
an open format, protected under some LGPL/GPL style licencing for word
processor files; a smart, generic, extensible, well designed file
format, supported by a good LGPL library, to encourge Linux word
processor authors to use this format; and similar things for other
application types.

Not an easy job, anyway, any taker :-> ?

Maurizio

-- 
Maurizio De Cecco                 Real Time System Team
				  IRCAM, Centre Georges Pompidou
                                  1, Place Stravinsky 75004 Paris, France

tel:   +33-(0)1-44784779  - fax:   +33-(0)1-44781540 - email: dececco@ircam.fr

Date: Mon, 05 Oct 1998 13:55:08 -0600
From: Jeffery Cann <jcann@fairway.com>
To: editor@lwn.net
Subject: Linus Torvalds

Hello.

I just read that the GNU project has been taking nominations for the
first Free Software Award (http://lwn.net/1998/1008/fsa-nominees.html).
Strangely, the name of Linus Torvalds was not on the list.  Surely this
was an oversight because without each other, the Linux kernel and the
GNU tools would not enjoy their current notoriety.  Is this a plot by
RMS because people won't call Linux 'GNU/Linux'?  I sure hope not.

Sincerely.
Jeffery C. Cann
jc_cann@ix.com.com

Date: Sat, 03 Oct 1998 04:29:45 +0100
From: teeth <teeth@fluffy.force9.net>
To: editor@lwn.net
Subject: Save the Xaw scroll bar.

I am driven to write by two things; Andrew Spencer's thoughtful look
at the RiscOS GUI and wide agreement at the recent EdLUG meeting.
  
  There seems to be a tendancy in all the developing Linux GUIs to
impersonate MS Windows, Motif and CDE even where there is already a
superior approach, the most glaring example being scrollbars. The
functionality of the Xaw/GNUEmacs scrollbar is far superior to the
MS/Motif approach, requiring, as it does, much less movement to use. It
is not perfect, scroll buttons would be a useful addition, especially if
there were both up and down buttons at each end (left and right too, for
that matter). 

  For Linux to gain desktop converts eye candy is not enough (great as
it is to raise interest), the usability must be superior in an
imediately accessable way.  I have yet to meet a windows user
unimpressed by the Xaw scrollbar, though the shine goes when it's rarity
is known. 

  Let the call go out "Save the Xaw scroll bar!" - may its children
flourish.


Alistair Murray

Date: Tue, 06 Oct 1998 17:01:08 +1000
From: Jeremy Lee <jeremyl@hrmc.spam-sucks.com.au>
To: editor@lwn.net
Subject: Acorn & Linux Revisited


Like Andrew Spencer (bspajs@bath.ac.uk) who wrote in last week, I have
also extensively used various Acorn RISC-OS machines. Indeed, I've still
got my elderly and battered Archimedes A310 (serial number 801!) stuck
in my cupboard. And the news that Acorn is dropping the Arc/RISC PC for
set tops is harsh to me.

I have just one suggestion to the Linux community, of which I consider
myself a part. Bring the Acorn users over! Now!

Any merger will be very hard. Arc coders are often fanatics who have
been defending their machines from MS/Windoze for as many as five years
before Linus even hacked his first kernel. And any suggestion that they
just roll over and be assimilated by Linux will be fiercely resisted by
those that matter. We need to offer the two related things that the Arc
hackers must surely be wanting: Hope, and A Future.

The reason I assert the Linux community should embrace the Acorn tribes
is because they have so much to offer us.

For a start, RISC-OS has always had an innovative and beautiful user
interface, perhaps the best I've ever used. Apps were consistently
written to the common UI guidelines, because they _made sense_. The Arc
had outline fonts, taskbar, inter-application drag-and-drop, pop-up
menus, and 'plug&play' in 1989. And if there's something the Linux
community needs right now, it's good interface designers with a history
of getting it right.

More importantly, the Arc was a doddle to configure. The Arc tribes will
hopefully bring an intolerance to messy configuration scripts and banish
forever such atrocities as sendmail.cf which seem merely 'quaint' to
Unix gurus.

Lastly, the arc is build with the ARM chipset. You've all heard of the
StrongARM? More powerfull than a Pentium II, while running on an AA
battery? The Arc embodies some fearsome processing power, and a talent
pool of people who know how to work with this architecture.

It's difficult to convey the depth of technical expertise which this
machine embodies, and the skill posessed by many Arc coders. David
Braben (who wrote Elite [yes, *the* Elite] Virus/Zarch, Conquerer and
others) was one famous early coder who made this machine do amazing
things. Because of the educational focus of Acorn, Arcs (like the BBC
micro before it) turned up in many schools and univeristies in the UK,
New Zealand, and here in Australia. People learned to code on this
elegant machine, and code well.

Acorn was also cool enough to sell and support a version of Unix for the
Arc for quite some time - A BSD variant, I think - so there are
considerable Unix skills already present. And the Archimedes was one of
the early port targets for Linux. (Though work has been a little slow)

Perhaps one of the reasons there's been less progress on the Arc port of
Linux is because RISC-OS is such a good operating system to begin with!
What it lacks in some modern features (eg. pre-emptive multitasking) it
makes up for in sheer speed and ease of programming. Half of the things
I've written for the Arc were in Assembly, because - hard as this is to
believe - it was often easier than using a high-level language. There
was a 'bare metal' feel to the machine which made it a Hacker's joy.

I personally feel a little betrayed by Acorn, and I've not seriously
used my Arc in years. To current gen, this will be a serious blow. We
are the new Amiga-nites, suddenly bereft of a homeland.. er.. supplier.
We'll coast for a year, until RISC-OS starts looking too dated. Then, if
Linux is ready for us, we'll come.

And Linux will gain another group of formidable, talented developers.

See to it.

--
.  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .
 Jeremy Lee | Orinoco              "One Crowded Hour of Glorious Life
 jeremyl@hrmc.spam-sucks.com.au       Is worth an age without a name."

    http://www.geocities.com/researchtriangle/lab/5003