See also: last week's Back page page.

Linux Links of the Week

WeWantHackers.com. A new job site is up: WeWantHackers.com. They are looking for high-quality developers, and have included a simple "hacking" test in the resume submission process to perform the first level of filtering. This site is run by Bruce Perens and the Linux Capital Group, so the jobs behind it could be interesting.

Section Editor: Jon Corbet

This week in history

A guy named Dave Whitinger announced the expansion of his "Threepoint Linux News" service. It would take a few more months yet for that service to evolve into LinuxToday.

The development kernel was 2.1.112 - and alleged to be in deep feature freeze. The 2.0.36 stable kernel release was in the prepatch stage. The beer-drinking penguin logo in the development series came under criticism - some people thought it would cause Linux to be taken less seriously. It eventually came out.

This was also a big week for Linux distributors; see this week's distributions page for a recap.

One year ago (July 29, 1999 LWN): The Netscape/Sun alliance backtracked and said that the Netscape Application Server would not be made available for Linux. Both IBM and VA Linux Systems announced plans to get into the Linux support business. Rumors went around that the (rumored) Transmeta processor would be used in the (rumored) new Amiga.

Life got difficult for those trying to participate in Red Hat's community stock offering, as the E*Trade system told them they were not eligible. Red Hat was breaking much new ground with this offering, and the financial system was having a hard time adapting.

No kernel releases happened this week; the stable kernel remained at 2.2.10 and development at 2.3.11. Stephen Tweedie's raw I/O patch was accepted, however, providing a long-missing functionality to the system.

Both the sourceXchange and CoSource.com started ramping up their operations this week. Bruce Perens' Technocrat also launched this week.

Letters to the editor

From: "Aaron J. Seigo" <aseigo@mountlinux.com>
To: letters@lwn.net
Subject: Miguel de Icaza: Unix and policy
Date: Thu, 20 Jul 2000 09:25:28 -0600

Hi.

The more I watch Helix Code step into the lime light under the reigns of Miguel
de Icaza the more I scratch my head. Does this man actually Get It(tm)?

The reason Unix does not enforce policy on the system level is because policy
belongs at the application level. Why? Well, look at BeOS/Mac/Windows. They
each enforce policies on programmers and users at the system level, but because
of that they are each cordonned off into their own space of the computing
arena. Unix is a substrate that strives to be flexible enough for _any_
policy. The idea is to OFFER choice, not limit them. As soon as one sets policy
of any sort of strictness, you immediately lose a good portion of your
usability (by definition). By keeping such policy decisions at the application
level, you can have the proper set of policies in force at the proper times and
in the proper places.

To say that a lack of policy creates a "defense system for hackers" is
to slap the Open Source community's reputation down and is just plain wrong.
Look at the amount of responsibility many open source programmers in the
UNIX/Linux world take upon themselves when it comes to security, performance
and feature set. In fact, this sounds more like a Redmond compaign than a
pro-Open Source developer's words.

I point to Icaza's own project Gnome as an example that he is (to
quoth him) "smoking crack" when spouting these arguments. Gnome sets policy, and
in the right place, too: on the application level. In fact, it is only because
of the gift of choice Unix brings that Gnome is possible; it is only because of
the gift of grace (aka "defense system for hackers") that Gnome was allowed to
stick around for as long as it did despite being as horrible as it was to come
to point where it has the opportunity to actually mature.

So I sit here and scratch my head. And ask: would I entrust my desktop to a
project whose leading visionary is so out into left field? Then I
remember: Unix doesn't force policy on me and limit my choices: I can use KDE,
or Blackbox, or CDE, or the command line, or.... and suddenly Icaza's rantings
seem distant and irrelevant. Which they are.
______________
Aaron J. Seigo
Mount Linux

Date: Thu, 20 Jul 2000 14:24:26 -0500
From: "Stoner, David M." <dmstoner@utmb.edu>
Subject: re: Unix sucks
To: "'letters@lwn.net'" <letters@lwn.net>


The editors:

Miguel is kidding, right?  If he really liked the Microsoft way so much,
presumably he would be working for Microsoft.

Here are what I hope are obvious retorts to his points as summarized in
your article:

>Unix is stagnant.

As a programmer, what I want above all else is an operating system that
never changes, or at least the API never changes.  I don't need a system
that keeps shifting underneath me and breaking my code.

>Unix's problems stem from ... not deciding policy.

The user/administrator decides policy.   Isn't that what we want?

>High level components

Fine if you want a bloated operating system which is much more than an
operating system; if you control the OS and mean to control everything else
and squash anyone who threatens you; if you assume that no one beyond your
control will ever write an application.

Yet there are high-level components for Unix, the use of which is
fortunately still optional.  For example, there is CPAN.  And there are
"glue" languages such as Tcl and its extensions.

Miguel is welcome to develop high-level components if he wishes, but I hope
he doesn't think they should be part of the operating system and everyone
should be forced to use them.

David Stoner
dms@atlantis.utmb.edu

Date: Thu, 20 Jul 2000 18:22:47 +0200
From: Martin Cracauer <cracauer@BIK-GmbH.DE>
To: letters@lwn.net
Subject: Staroffice GPL

Most people cheering after the annoucement overlook that they won't
give us the source for the currently working product.

They give us the source for some future experimental stuff.

That will be the same situation as with Netscape/Mozilla.  We aren't
allowed to fix our lovely Navigator-3.x (which is what most people I
know would prefer), we have to hack on some Uebercode that in any case
is slow and probably won't lead anywhere in the foreseable future.

In no way I can share LWN's editor's opinion that Sun tries to avoid
Netscape's mistakes.  They do exactly the same thing, the delay
between annoucement and release (why not next weekend?), the refusal
to give out the working cod and even the XML-based components
approach.

Not a problem for Sun, of course, it will be "OpenSource" that has
failed.

-- 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Martin Cracauer <cracauer@bik-gmbh.de> http://www.bik-gmbh.de/~cracauer/
BIK - Aschpurwis + Behrens GmbH, Hamburg/Germany
Tel.: +49 40 414787 -12, Fax. +49 40 414787 -15

Date: Thu, 20 Jul 2000 20:56:46 +0300 (EETDST)
From: Tuomas Lukka <lukka@math.jyu.fi>
To: letters@lwn.net
Subject: GPL StarOffice

Let's not forget for a moment that there's only one reason for Sun to
release StarOffice under GPL: they want to hit Microsoft where it really
hurts. And I believe that it will; this is what is needed to topple their
empire. Say in about a year it will be possible to provide a cheap,
reliable, care-free distro that can do all the things that business users
want and is completely open-source. Microsoft does NOT want to see that
day.

So while the announcement is great news, don't make the mistake of
trusting Sun any more because of this. We have a common enemy but once
Microsoft crumbles, I don't expect any more such actions from Sun. Which
is all right: GPL will stay GPL, and companies need to compete.

But just remember: even though it isn't really said publicly anywhere
(the FAQ has some wording which alludes this but Microsoft is not
mentioned), this move takes direct aim at Microsoft. Nothing more, nothing
less.

	Tuomas

Date: Thu, 20 Jul 2000 11:17:28 -0700 (PDT)
From: "Robert A. Knop Jr." <rknop@lilys.lbl.gov>
To: letters@lwn.net
Subject: The source of animated images on web pages

I am finding myself increasingly annoyed by animated GIFs on web pages.  
I think this is a result of the fact that more and more of the pages one
visit have banner adds, or sidebar ads.  As much as I hate to say it, the
site that inspired this letter was LWN.net itself.  On LWN, as with many
other sites, I find it hard to read the text (i.e. the whole reason I went
to the web page in the first place) when, at the corner of my eye, there
are these things flashing and drawing my attention away.  Advertisers will
be very pleased to hear this, but I find it irritating and annoying.

Obviously, one will say that I should just hit ESC or the "Stop
Animations" menu item.  Unfortuantely, in Netscape (the only option for a
Linux web browser at the moment, if you care about things like https and
Java/JavaScript), ESC is also the "stop loading" key.  I want is to stop
the animated GIFs from flashing, I don't want to stop anything on the page
from loading.  It seems impossible to stop the images from flashing until
everything on the page has loaded. Netscape 4.73 doesn't even seem to be
entirely consistent about when the "Stop Animations" menu item is
available (via the right mouse button).  Especially if the page has
borderless frames, sometimes it seems to be nearly impossible without a
lot of work to get the animated images to stop.  This does not make for a
pleasing web browsing experience.

The ideal solution would be an international agreement for everyone to
stop using animated images on web pages.  This is, of course, completely
unrealistic, and I don't even expect LWN.net to do it.  But, for future
developers of web browsers, I would ask that you have a key or icon that
will stop all animations on the current page, even if that page isn't
fully loaded yet.  What's more, a preferences option to disable animated
gifs altogheter (with perhaps an override key to start the animations
again) would be really nice.  (This is analogous to the option many
browsers have of just not loading the images unless you push the "images"
button.)

If Netscape already has an option like this, somebody please let me know.
I haven't been able to find it.

-Rob Knop
rknop@lbl.gov

Date: Thu, 20 Jul 2000 16:54:03 -0700
To: letters@lwn.net
Subject: RSA Patent Expiration
From: Rick Moen <rick@linuxmafia.com>

Dear Mr. Corbet and Ms. Coolbaugh:

You may be amused to hear that the data and time the (USA) patent on RSA
expires may be literally indeterminate!  

The RSA algorithm is covered by US patent #4,405,829, which was issued
September 20, 1983 (and thus is a 17-year patent) to Ronald Rivest, Avi
Shamir, and Leonard Adleman.  Everyone says it therefore expires this
September 20.  Here in the Bay Area, we have tentative plans for a
celebration party and informational event concerning RSA and other
crypto software, when the patent expires.  But, the question is, when
exactly?  5 PM Washington DC time on the 20th?  8 AM Washington DC time
on the 21st?

Hoping to find out when exactly the algorithm can be freely used in the
USA without royalties or infringement, we posed that question to some
friendly and knowledgeable patent attorneys at the firm Bever, Hoffman,
and Harms, LLC (http://www.beverlaw.com/), in San Jose.  Partner Julie
Stephenson kindly responded, with an e-mail from that firm (which
presumably should not be considered legal advice):

---<snip>---

I researched the question right after we spoke.  Unfortunately, the
answer is that there is no answer.  According to Chisum (a premiere
researcher in the field), the caselaw on the subject is in conflict.
Thus, if a patent has a date of June 28, 1983, and the term of the
patent is 17 years, then the last day of coverage of the patent has been
interpreted to be both June 27, 2000, and June 28, 2000, in different
cases.

I looked around a little further, and found no information relating to
the time of expiration of a patent.  Because the caselaw is still in
conflict on the date of expiration, I can't imagine a situation where
the *time* on the date of expiration (much less the time zone of the
time on the date of expiration) would have been litigated without
clarifying that whole date of expiration mess.  So I can give you no
direction as to *when* you should begin partying.

What does this mean for you?  Well, you can either party on September
20, 2000, and be prepared to change the name of your party from "the
first day of no coverage by the RSA patent" to "the last day of
oppression by the RSA patent" while knowing that anyone partying from
11:55 pm to 12:05 am will have actually partied on the right day
(ignoring that whole time zone thing) OR you can party on September 21,
2000, and be assured that you are partying on a no-patent coverage day.
However, in the minds of some people, you would be partying on the day
after the day the patent expires.  :)  If it helps a all, it appears
that generic drug manufacturers would begin selling their drugs on (in
the example above) June 29, 2000.  (Note that this ignores the issue of
then manufacturing the drugs prior to the expiration of the patent,
which was one of the bases for litigation in one of the conflicting
cases mentioned above.)

Sorry I couldn't be more clear - that is the pitfall of working in the
law.... There is often no right answer, only opinions and arguments.  On
the bright side, you can choose a reasonable time and date of expiration
(say, 11:59 pm EST on September 20, 2000), and have some caselaw basis
for choosing that date.  If you think of it, please let me know what you
decide.  I'll have a drink at that time in celebration.  :)

Have fun,

Julie

---<snip>---

Happily, after one notices that the 20th is a Thursday, the alternative
of skirting all these issues by holding the party on Saturday the 23rd,
instead, became (er...) patently obvious.

-- 
Cheers,                              "Open your present...."
Rick Moen                            "No, you open your present...."
rick (at) linuxmafia.com             Kaczinski Christmas.
               --  Unabomber Haiku Contest, CyberLaw mailing list