See also: last week's Back page page.

Linux links of the week

Other aspiring portal builders may want to check out the Linux portal mini-HOWTO on Humorix.

The Linux hardware database maintains a list of available hardware, along with ratings of how well each item works with Linux. It seems a bit thin at the moment, but if more people head on over and contribute their experience, it should easily develop into a highly useful resource.

Section Editor: Jon Corbet

Letters to the editor

Date: Wed, 09 Jun 1999 22:57:01 -0400
From: William Hoffman <whoffman@nospam.erols.com>
To: lwn@lwn.net
Subject: Sadly, you are wrong

about RedHat's $80 price tag. For new users such as myself--sick of MS's
buggy bloatware, actively seeking a _productive_ alternative, ambivalent
in the face of a bewildering variety of near-commodity type Linux
distributions lining MicroCenter store shelves--the issue breaks down
like this:

Slackware: $20; too bad there's no manual
Debian: $30; yeah, right
SuSE: $40; intriguing... but I don't read German
Caldera: $40; boasts an easy install (and with MicroCenter's 30-day
money back guarantee, I can experiment)
Red Hat: $80; compared to the others... pricey

It's immaterial that these impressions are flawed (SuSE's fully
English-language manual is one of the best, though the hard-to-pronounce
name gives newbies a fit). Only propellerheads will appreciate the
support Red Hat has given the Linux community, or care about the
controversies the company's strategies have provoked. To paraphrase an
early 20th-century president of these United States, "What this country
needs is a good $50 operating system." In other words, the future
belongs to Corel, if they know how to take it.

William Hoffman
 whoffman@nospam.erols.com

From: "Greg Mader" <gmader@geoanalytics.com>
To: <letters@lwn.net>
Subject: smbmount woes.
Date: Thu, 10 Jun 1999 09:24:30 -0500

Editor,

For way too long, smbmount has performed poorly. The Samba team has
washed its hands of it, saying "We are working on smbsh," and
"SMBmount is part of smbfs, not samba."  Well, the corruption problem
in today's Linux Weekly News is the last straw.  The Samba team has
done a marvelous job creating a one way server system. Linux does not
have high quality SMB clients now, and Linux probably will not have
any for some time. I strongly urge the Samba team to "put up or shut
up" on smbsh, or start working on smbfs to become stable, safe, and
full featured. In the mean time, I will suggest Sharity to people who
ask for connectivity from Linux to NT servers, and warn them that in
my experience, SMBmount is a dud.

Greg Mader

Date: Mon, 14 Jun 1999 18:28:26 +0200
From: Alessandro Muzzetta <muzzetta@geocities.com>
To: letters@lwn.net
Subject: Danger of software patents in Europe

Hello,

I'd like to bring to your attention an article by Richard Stallman
warning on the danger of software patents in Europe:

	http://features.linuxtoday.com/stories/5960.html

The article may be copied and redistributed verbatim.

Stallman has been notifying European users of the imminent introduction
of software patents in Europe by the European Parliament, before
Summer 1999.
As you probably know already, software patents are considered a 
Bad Thing, especially for the Free/Open Source software community.
The European Union intends to emulate the flawed US patent system.
Thus, software patents that are making life hard for US programmers
today (consider the GIF patent, the RSA patent restricting PGP, etc..)
may find their way into Europe.

There is already a group in France (www.freepatents.org) that is
involved in the fight against software patents.  Similar initiatives
have been instituted in Germany (swpat.ffii.org) and Italy
(no-patents.prosa.it). 

It seems people are uninformed.  We are desparately trying to raise
awareness on the issue and would appreciate your help in doing so.

Thanks for your cooperation.


Alessandro Muzzetta

Date: Mon, 14 Jun 1999 18:26:47 +0100
From: Richard Kay <Rich_Kay@ect.uce.ac.uk>
To: jp@ncfocus.com, letters@lwn.net
Subject: Look before you leap

Dear Mr Morgenthal,

Concerning your article in Internet Week online this gives the
impression that the silly season has started in earnest this year. 
Have you actually used a recent version of Linux ? Much of the 
positive press this development has been getting has until recently
been due to the very rapid pace of development of this system, based
as it is on open collaboration and peer review - something
the scientific community has found more effective
than the alternatives for many years in other areas. However
more recently this system has been able to stand up to more
serious comparison. This is not to say that Linux is best in
all areas yet. But Sun and IBM know well enough how to read
the writing on the wall and which side their bread is buttered 
as they receive somewhat more revenue from services than sale of 
their own proprietary operating software.

You state:
                   
>Linux is an open-source project; therefore, all changes to the kernel
>are subject to review and approval by a small team that controls
>this portion of the operating system. Companies that add features
>they need, but that are not accepted into the core distribution, may
>find themselves in a redevelopment and retesting cycle every time
>a new version of Linux is released.

Firstly only a few users need changes to the kernel and
those who do are less dependent on the core kernel team if they use
Linux than if they use a system to which they have no direct
access to read and modify the source code. Secondly those who want 
to add features are free to decide their own upgrade cycle and
branch the kernel development if they so wish. I don't imagine for
example that those using Linux for embedded systems will ever want
to incorporate all of the mainstream kernel into their subset/branch 
of it. Forthly the reason there has been so little branching 
(attested by Microsoft's Halloween documents) is testimony to
the success of the core development team in incorporating features
needed by the development/user community. The few who create
and use dependencies through binary-only patches without 
providing access to source code have only themselves to blame
if kernel changes make life slightly more difficult for them. 

The fact that desktop Windows users still outnumber desktop Linux 
users is somewhat irrelevant to those considering 
applications for which Linux is better suited, e.g. web servers
where Linux servers have more web client users now than NT does.

For large SMP systems Solaris and AIX will not dissapear 
overnight - but as has been said before IBM and SUN are not 
supporting Linux for reasons contrary to their own commercial 
interests.

On the complexity front, yes there is still a problem with
Linux as it is a fast growing one with NT. But it is better to 
build a complex system upon a sound design than a shaky one.
Also an increasing range of Linux system administration facilities
are being implemented using simple web front ends and 
Windows-style GUIs and Wizards etc. The fact of Linux's
low cost is partly instrumental in its fast growth of use
within my own field of higher education. The inevitable effect of 
this will be to ensure a very rapidly growing stream of people coming 
out of education with the skills needed to administrate, manage
and configure this system.

In light of its success as an Internet/Intranet server platform and
development in other areas and the recognition of their need to support 
this system by IBM, HP, SUN, Oracle, Informix, Dell and Compaq etc, your
calling it a "a college student's project gone astray" is just plain
daft. 

Hence my question about whether you are writing about something
which you have ever used ? 

Yours Sincerely,
Richard Kay
School of Electronic and Software Engineering,
Faculty of Engineering,
University of Central England
Perry Barr,
Birmingham B42 2SU
UK
Email: Richard.Kay@uce.ac.uk

To: letters@lwn.net
Subject: Justin King on osOpinion...
Date: Wed, 16 Jun 1999 16:50:15 +0100
From: kevin lyda <kevin@suberic.net>


Just read Justin King's piece on osOpinion.  I hope no one mailed him
to enlighten him; it would be like sending an interior designer into a
black hole to "brighten up the place."

I'd love it if the "Linux is just a craze" crowd would also let us know
what they thought about the net five years ago.  These people wouldn't
get a clue if they were caught in some freak clue blizzard, never mind
a few emails.

Unrelated to Mr. King's article, I'd also like the "Linux thing is just
a new version of ABM" to explain just when they first heard of unix,
and if they're aware that 1969 came just a few years before 1980.  Opinion
pieces are nice, but without any meaningful facts they should be more
accurately labeled as ravings.

Thanks,

Kevin

Date: Thu, 10 Jun 1999 16:32:11 -0700 (PDT)
To: letters@lwn.net
Subject: *code wars* [Qt reserved words]
From: Leo Razoumov  <lrazoumov@qualcomm.com>

Nathan Myers brought a very valid point of a library design
emphasizing the fact that Qt claims several general-purpose tokens for
Qt library.  I do digital signal processing and telecommunications
software development where 'signal' and 'slot' have their natural
roots and meanings. This is why I do *NOT* use Qt.  By now C++ has
been around for a decade or so and people already worked out simple
rules how to avoid global namespace pollution by prefixing the
keywords:

String  => RWCString  (Rogue Wave Class String)
Display => XtDisplay  (Xt Intrinsics library function)
signal  => QTsignal   (proposed for Qt library) !!!!

I think it is just a good taste in library design to either use
namespaces as defined in C++ ISO standard or to stick with library
prefixes for a while.  I personally would not complain to type extra
QT in front of everything which comes from their library.

On the other hand the Nathan's appeal to deliberately break Qt based
applications sounds strange (to put it mildly). So far Troll Tech has
been receptive enough to address the needs of Open Source
Community. Also they do not want a negative publicity for their
flagship commercial product when people start questioning the quality
of their library design.  I sincerely beleive that the issue could be
solved in timely manner and without drastic measures like *code-wars*.


thanks,
-- Leo Razoumov, Ph.D.

DISCLAIMER: Statements and opinions expressed in this message are the
private optinions of its author and do not represent or reflect
explicitly or implicitly opinions and positions of the author's
employer.