See also: last week's Back page page.

Linux links of the week

Stokely Consulting's Unix Sysadm Resources is a comprehensive site full of good stuff for administrators of any Unix-like system, including Linux.

Letters to the editor

Date: Fri, 5 Feb 1999 02:02:28 -0500 (EST)
From: cph@martigny.ai.mit.edu (Chris Hanson)
To: editor@lwn.net
Subject: Structure vs purism

OK, I see that someone else has chosen to speak up about gotos in the
kernel, so I'll chip in too.  I agree that more context should have
been given.

Where I disagree, and very strongly at that, is in the implication
that goto's are somehow "unprofessional".  Goto is simply a tool, and
its use can be either beneficial or harmful, depending on the
programmer.  A skilled programmer uses language constructs to provide
a clear expression of the program; goto is appropriate in many
circumstances because it captures the intent of the program more
accurately than other available constructs.

Goto got an undeserved bad rap from the structured programming
movement of the 70s, which was epitomized by Pascal.  Pascal lacked a
goto statement, which meant that if the control structure of a program
was not cleanly expressed by the available iteration constructions,
you were SOL.  I've written a fair amount of code in Pascal, and I can
say from experience that I missed the goto statement when programming
in that language; at least C _has_ a goto statement, and I use it when
I think it is appropriate.

It's also kind of funny that this conversation comes up in the context
of Linux.  To me, one of the strengths of the Linux community is its
attention to freedom.  Goto was originally denigrated by people who
decided that the way to prevent bad programmers from writing bad
programs was to reduce the expressive power of the languages --
specifically by eliminating goto.  (Of course, bad programmers will
_still_ write bad programs, even without goto.)  Now why would someone
who believes in freedom want to _reduce_ expressive power?

Now, you may be wondering why I feel so strongly about this.  The
answer is that I'm a Scheme programmer.  In Scheme, there is only one
iterative control construct: the tail-recursive procedure call, which
is just a goto with arguments.  So by denigrating goto, you indirectly
denigrate my favorite language, in which I write many beautiful and
elegant programs, each filled to the brim with "gotos".

So please, have a little respect for goto, and those of us who like to
use it.

Date: Mon, 08 Feb 1999 11:23:58 +0000
From: Mark Lamb <dragon@freedom.org>
To: lwn@lwn.net
Subject: A Plea for help

Finally managed to carve out some time to play with 2.2; am now trying
to wrap brain around new routing code in particular. The ipchains and
ipmasqadm have decent man pages; the iproute2+tc stuff has only BNF-ish
descriptions of the syntax.

Anybody got any idea how to use any of these bright shiny new toys? I'm
taking hints, for (possibly delayed) publication, via email at 

	dragon@snafu.freedom.org

I'm hoping to write some coherent docs; until I get the time I'll post
anything sent raw for the good it might do others. It'll all be at 

	http://snafu.freedom.org/linux2.2/

Along with copies of some of the packages I've downloaded to get a
RedHat 5.2 system ready for 2.2 et al. This includes the latest iproute2
package (as of Jan 23) from
	ftp://ftp.inr.ac.ru/ip-routing/
Which seems to have been down for the last several days.


-- dragon@snafu.freedom.org (Mark Lamb)
I won't cry for the wasted years cuz' you ain't worth the salt in my
 tears.

Date: Wed, 10 Feb 1999 22:40:43 -0700
From: Jeffery Cann <linuxguy@ix.netcom.com>
To: editor@lwn.net
Subject: Zope License

(Slightly modified version sent to Digital Creations)

Greetings.

As I see a Zope advertisement on the LWN site, I am concerned that the
Zope development environment is "Open Source" branded.  Specifically,
the third clause in their latest license version (0.9.7) continues to
troubles me.

Initially, Digital Creations (www.zope.org) wished to require
attribution when web sites were produced using Zope.  More recently,
they have come closer to freeing their code.  IMHO, the point of section
3 is still works against the rationale of freeing software, whether
"open source" or via the GNU Public License (or similar licenses).  The
point of releasing source code is so that the community may share it,
improve it, test it, etc.  The benefits received by Digital Creations
will outweigh any concerns with attribution.

Let me draw on the Apache web server (and license) as an example, since
the Zope Public License (ZPL) is based on it.  The Apache web server is
the most popular and most used web server in the world.  It became so
because their software was continuously improved by a
community of developers.  Eventually, it overtook the reliability of
competing proprietary products, such as Microsoft's IIS or Netscape's
web servers.

Apache does not require or even ask that attribution be given to the
Apache development group.  Yet, amazingly, every technical person who
develops web content knows that Apache is the web server of choice.  The
reason every one knows this is because Apache is technically superior
server.  It is the most popular due to its technical merits, not because
they required or pleaded for attribution.

If Zope is a real alternative to Cold Fusion, then why still ask for
attribution?  Why not let the product stand on its technical merits?  I
guarantee you that if Zope is better or equal to Cold Fusion, it won't
take too long for the web developer community to figure this out and
free attribution will follow.  It didn't take long for Apache to become
the number one server.

Finally, I must quote the Chinese philosopher Lao Tsu  -- "Who does not
trust enough will not be trusted."  Freeing source code is about trust
-- trust between the person(s) or company who wrote the original code
and those in the community who use it and contribute it.

I appreciate the efforts of Digital Creations to refine the ZPL.
Hopefully, they will drop clause 3 altogether so I can start using the
Zope development environment.  Until then, I am happy to continue to
hack perl, gimp and htmlpp for my web site development efforts.

Sincerely,
Jeffery Cann