Mozillazine has a new Firebird preview article: (Rapid Pace of Development for Mozilla Firebird. Among other things it addresses one of the issues Anil brought up about accessing sidebars.

First of all, any bookmarked page can be made to load in the
sidebar via a simple checkbox in its properties dialogue. In addition,
support is also offered for the Mozilla Application Suite’s addPanel API,
which causes a bookmark that is already set in to open in the sidebar
to be created. Furthermore, Opera’s method for adding panels to its
sidebar equivalent, the Hotlist, is also supported (this method simply
involves a normal link with a rel="sidebar" attribute). By creating a link with a target="_search"
attribute (named for compatibility with Internet Explorer), a site can
cause a page to be loaded transiently in the sidebar, with no bookmark
being created.

Very good.

eZ publish 3.1: Take It For A Spin

Pretendster, generates fake friendsters and testimonals. Care of Tree-Axis and Fake I.D..

• Optimizing JavaScript for Download Speed
• yp’s JS Kompressor
• JavaScript Crunchinator
• ESC
• Javascript Copmpression
• Extended ACSII Javascript Packer

todo: wiki comparison

criteria: code modularity, data model extensibility: how to integrate blogging, todo (simple node states?), development activity, diffing/versioning, authZ model

• Twiki
• Tavi
• Kwiki
• TikiWiki
• PHPWiki
• CoWiki
• SnipSnap

wolrahnaes does a Opteron vs G5 system price comparison

A WikiWhiteboard with SVG would be perfect for mapping out stuff. Actually, having relational diagramming primitives and multiple layers would be perfect, but beggars can’t be choosers (one step at a time). via paranoidfish

I started noticing the worm spoofs this morning around 10AM. Doing some legwork seemed to reveal it as a Sobig variant, and about a half hour later the virus definition popped up in the AV scanner. Around the net, a lot of people are clarifying that this isn’t an “email worm,” but a “Windows worm,” which in some sense is true, however, while the worm only propagates through infected W32 machines, the spoofings/spoofed bounces affect (and highlight the problems) of the entire email system, and highlight the need to some sort of authN/dsig system.

Perhaps signed headers would be a simple way of solving the problem – it’d require some extra key servers, but you could implement this on a per-server basis, or even the user level (mta level would be better, it would save bandwidth and allow realms of trust), and is completely backwards compatible…

(While perhaps resource intensive, it would certainly create a trusted path, and could be done completely voluntarily (ie, deployed when the cost of sifting/sending bad emails outweighs the cost of the decrypting the signatures of each message).

todo: look up proposals for various systems/frameworks that can verify paths, senders, recipients, but still maintain some semblance of privacy or anonymity

• LWN: Why SCO won’t show the code

  Comment by mdrejhon:

  Found the V5 source code in its entirety:

  Photo of SCO Slide:
  http://www.heise.de/newsticker/data/jk-19.08.03-000/imh1.jpg

  Near exact match to 30 year old source code in V5:
  http://minnie.tuhs.org/UnixTree/V5/usr/sys/ken/malloc.c.html

  What the heck is SCO doing, claiming 30 year old source code as their own?

  Interesting, this code was removed 7 weeks ago because it was “ugly as hell”:

  There’s also a very informative lkml thread [theaimsgroup.com] about this and it’s already been removed from the source tree [bkbits.net], but apparently not because of copyright issues but because it was just “ugly [theaimsgroup.com] as hell [theaimsgroup.com]”.

• Analysis of Linux Code that SCO Alleges Is In Violation Of Their Copyright and Trade Secrets. – Bruce Perens
• TWikIWeThey: SCOvsIBM
• SCO’s now claiming over a million lines of offending code (Linux 2.4.20 is ~4.4M LOC) – that they can’t show anybody, um.. because it’s a big secret, yeah… that’s the ticket

  Insight from FJ:

  Perhaps a more informative estimate would be to take the difference in the number of lines in the 2.2 kernel vs the 2.4 kernel, since SCO claims 2.2 is fine but 2.4 has copied code.

  3,377,902 (2.4) – 1,800,847 (2.2) = 1,577,055

  In other words, SCO claims that 2/3 of the improvements in the 2.4 series kernel belong to them. That is a rather unrealistic statement since a lot of those enhancements didn’t come from IBM.

  Also:

  The SMP code (all platforms) seems to be ~15000 lines of which ~800000 lines is copied verbatim from sysV.

  So, am I to assume that the Linux kernel is 5333% infringing code?

  Man, I wish I could code that efficiently.

Hey, so if SCO claims the GPL is invalid, by what license are they distributing Samba in their new version of OpenServer?

The SCO Forum crowd applauded when SCO executives announced that an
upcoming version of its OpenServer–code-named Legend–will support the
latest releases of Java; include new hardware support, such as
universal serial bus (USB)
printer drivers; contain expanded security features; and provide better
compatibility with Microsoft Windows through version 3 of Samba,
which is developed by an open-source group. The OpenServer update is
scheduled to debut in the fourth quarter of next year.

This weekend and next week will involve trying to squeeze as much productivity out as possible; after that, I’ll be slightly occupied.