Advogato's Number

Posted 9 Nov 1999 at 08:00 UTC by advogato Share This

This is the first in a series of editorials by Advogato. Topics this week include the DVD crack, ReiserFS, and the Debian freeze.

This week, the encryption of DVD players was finally cracked. I say finally, because the code was pitifully weak, even by 40 bit encryption standards. I came across a block diagram of the encryption algorithm on a whiteboard in Berkeley, and it was obvious even to this cat why it sucked.

This event has one happy implication for free software: it makes a free software player feasible. In fact, a very early prototype of such a player exists already. With the talent and motivation of bright Linux programmers, there's little doubt that a nice, polished player will be available soon.

A more disturbing note, however, is the legal threat to the people who did the reverse engineering. There are already enough laws that have been effectively used to block free software development (don't get this cat started on export restrictions). The Japan-based DVD forum apparently has gone after Derek Fawcus first based on a law against distributing software that can be used to defeat copy protection systems.

Most of the reverse engineering work was actually done by the Norwegian Masters of Reverse Engineering group. Reverse engineering is fully legal in Norway, so these developers have not (yet) been targeted.

Many have noted that the cat is out of the bag on how to defeat the copy protection code. While this is true, and I'm very happy about it (meow), it doesn't guarantee smooth sailing ahead for developers of free DVD software. The legal threat makes it almost certain that major distros won't be including software any time soon. And the decision to go after Derek will no doubt send a chilling message to other free software developers. Whit Diffie has spoken of the "marijuana-ization" of crypto (ie, you can get it if you want it, a significant minority does, but it's not officially sanctioned so most "normal" people just don't bother). It' s not hard to imagine that various intellectual property protection schemes will be treated the same way.

A wave of the paw and best wishes to Derek and all the people involved in the Livid project to bring free DVD decoders to your desktop. Follow the discussion on the Livid mailing list archives and of course Slashdot.

The lack of a journalling filesystem has long been cited as one of the things holding Linux back in mission-critical applications. This won't stay the case for long: not one but three such filesystems are in the works: Stephen Tweedie's ext3, the XFS code donated by SGI, and Reiserfs by Hans Reiser. And this week, Reiserfs went into real release, so there is now a journalling filesystem.

Reiserfs is interesting in a lot of ways other than the journalling, in particular the use of trees for both the metadata (ie directories) and the files themselves. Thus, Reiserfs is expected to do fairly well in a number of situations where a more traditional filesystem does poorly: lots of small files, a large number of files in a directory, and so on.

But one particularly intruguing aspect to Reiserfs is its development model, which is paid exceptions to the GPL license. Stig Hackvan, in his Linuxworld piece, called this the "pay to get out of jail" model.

It will be very interesting to see how this model works out. Traditionally, it's been used for fairly well defined components that are technically difficult to replicate, such as Sleepycat's handling of the Berkeley database. A filesystem, on the other hand, tends to be an integral part of the operating system. As a case in point, integration of SGI's XFS into the Linux kernel seems to be quite a bit harder than many people hoped.

Even so, the "pay to get out of jail" model promises to be a good one for both free software and its developers. Unlike the "pay for support" model, it actually rewards free software developers for the work they do best, programming. The existence of funding sources can enable free software developers the luxury to work on their projects fulltime, and really polish them. And, even though the code is licensed under proprietary licenses, it still exists in a fully free branch. To me, this approach strikes a fine balance of power - if the original developer mishandles the code, either technically or on the business side, other disgruntled colleagues can always create a fork. Yet, a developer continuing to manage the project well retains the control needed to get a decent licensing revenue. Advogato will watch Hans's career with great interest.

Finally this week, the Debian project announced that it was slipping its code freeze for the Potato release to at least Jan 15. By the standards of the rest of the Linux distros, this is ridiculously slow - a number Slashdot commentators mused that the 2.4 kernel may arrive before a stable Debian release with a 2.2 series kernel.

However, taking the time to get things right is one of the great free software traditions. Rushing things out the door to reach the market in time is a hallmark of proprietary software. Advogato commends Debian for holding true to their ideals. And while it looks like Debian might get passed by in the high-stakes distro market, it is quietly racking up its share of successes, including eBay choosing to run their servers on Debian, and of course Corel's choice of Debian as the basis for their work.

Advogato wishes the Debian team well, and reminds people that, while taking time to get things right is a fine goal, slowness is not in itself always a good thing, and development goes faster with competent help.

Until next week,


New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

Share this page