Not to be missed: well-known enfant terrible James Grimmelmann is guest-blogging on PrawfsBlawg. His opening commentary on the relationship between law practice and computer science:

Practicing lawyers, like practicing programmers, are professional pragmatists. Both must make their cases (and case mods) out of the materials they have available; both starve or eat steak depending on whether their creations work. The day-to-day practice of law is unlikely ever to require much high theory. We can mourn that fact because it means that they look at us with suspicion, or celebrate it because it frees us to chase Truth and Beauty and it will remain a fact either way.

Aside from the fact that I don’t eat steak, I think this is correct. Via a commenter on James’ entry, I learned that the 7th Circuit Court of Appeals is implementing a wiki (the entry page could surely use some more content). Surprisingly, it was not Posner but Easterbrook who spearheaded the effort. This is a very interesting development, but I expect it will be quite a while before any other circuit takes up the idea. Finally, I have been meaning to write about this New York Times story describing Jonathan Coulton’s success as a musician breaking with the traditional distribution /promotional channels (via 43 folders, a productivity blog that is still on my “probation” list). Unfortunately, slashdot beat me to it. I first re-discovered Jonathan Coulton during his guest episode of the Show with Ze Frank. In any event, the article is well worth reading:

More than 3,000 people, on average, were visiting his site every day, and his most popular songs were being downloaded as many as 500,000 times; he was making what he described as a reasonable middle-class living between $3,000 and $5,000 a month by selling CDs and digital downloads of his work on iTunes and on his own site Coulton realized he could simply poll his existing online audience members, find out where they lived and stage a tactical strike on any town with more than 100 fans, the point at which he d be likely to make $1,000 for a concert. It is a flash-mob approach to touring: he parachutes into out-of-the-way towns like Ardmore, Pa., where he recently played to a sold-out club of 140 . In total, 41 percent of Coulton s income is from digital-music sales, three-quarters of which are sold directly off his own Web site. Another 29 percent of his income is from CD sales; 18 percent is from ticket sales for his live shows. The final 11 percent comes from T-shirts, often bought online

Share This

More than a year ago, I published an election methods library called RubyVote. Interest in the library surpassed any of my expectations: I know of at least one startup using the library heavily in their core business and a number of fun sites, like Red Blue Smackdown, that are using it as well. The point of course, was to make complex but superior election methods accessible in all sorts of places where people were making decisions suboptimally. It its own small way, it seems to have succeeded enormously. Over the last year, I've been asked by a variety of people if they could use RubyVote for their own organizational decision making -- tasks like electing leadership of a student group or members of a non-profit board of directors. Since RubyVote was just a library without a UI of its own, I had to tell them "no." I caved in eventually and got to work on a quick and dirty web-based front end to the library. That project grew into Selectricity which is a primarily web-based interface to a variety of different election methods and voting technologies. You can currently try out quickvotes which can be created in half a minute and voted on in a quarter but which bring all of the power of preferential voting technologies to bear on very simple decisions. Prompted by Aaron Swartz, I also built a mobile phone version that's lets you send a short email or SMS to create or vote in a election. For those that follow research in voting technologies, there's not a lot of new stuff here. What's new is that this project, unlike the vast majority of voting technologies, is interested in the state of the art for everyone but governments. Clearly government decisions are important but they're one set of decisions, usually only once a year. Selectricity is voting machinery for everything and everyone else. It was announced in a variety of news outlets today that Selectricity was selected for grant from mtvU and Cisco as part of their Digital Incubator project. As part of that, I'm going to be working with some other voting technology experts to bring tools for auditable elections, cryptographically secured anonymity, and voter verifiability to the platform (I have only rudimentary functionality today). There are a couple people who will be joining me on the project this summer and we will building out what I hope will be an extremely attractive platform for better every-day decision-making. More than the grant though, I'm excited about the visibility that use by MTV will bring to the project. Most of all though, I'm just excited about more free software and more (and more accessible) democratic decision making. My adviser Chris Csikszentmih lyi put it well:

One of the big arguments against preferential voting, or new voting technologies, is the fear that they would disenfranchise the average person who doesn't yet understand how they work. Certainly, making all voting technologies open source is critical, but the issue of familiarity is worth considering. We re hoping that MTV and eventually American Idol will move their voting over to Selectricity, allowing it to work as both a technical tool but also pedagogically, training future voters. Why not integrate democratic processes into all your software and communications tools? Why not use the best democratic processes available, so long as they're available to everyone?

Currently I am considering the priority scheme to use for some highly available services running on Linux with Heartbeat. The Heartbeat system has a number of factors that can be used to determine the weight for running a particular service on a given node. One is the connectivity to other systems determined by ping (every system that is pingable can add a value to the score), one is the number of failures (every failure deducts a value from the total score), one is the weight for staying on the same node (IE if the situation changes and the current node is not the ideal node you might not want to immediately move the service to a different node as that gives some seconds of no service), and one is the preference for each node that may run the service. For a given node to run a particular service then the score has to be greater than all other nodes and also greater than zero. If all nodes have a score that is zero or less then the service will not run. Now in the case of a service that repeatedly fails (EG a filesystem mount that relies on a hardware RAID which is not connected) then what should we do? One option is to have the score for running on a particular node be for example 100 times the value that is subtracted on failure. In this case after 100 failures on that node (and an appropriate number of failures on other nodes which are permitted to run the service) it will be disabled. Then the service has to be explicitly re-enabled (or a node rebooted) before it will run again. The other option would be to have the value that is subtracted on failure be less than a billionth of the score for running on a particular node, so that the service will keep trying to start for the next few hundred years. The up-side of this is that there is less fiddling required, the down-side is that some CPU and disk resources will be kept active in repeatedly starting the service. Now I have to decide which option to take in this regard, any comments would be appreciated.

Yay! Yet another release party! See you there!

While the announcement from ATI about their intention to provide better support for Linux is certainly welcome it’s not clear what exactly they intend to do. They don’t appear to have actually said that they will release their drivers. This could actually be a really good thing - for example, I’d expect that documentation and support for people producing free drivers would be at least as useful as their current proprietary code - but it’s worth bearing in mind.

Today is the end of a pretty long week-end of dealing with samba. About 10 days ago, we (the samba packaging team in Debian) were privately notified of security issues found by the Samba Team developers in this quite popular package. This very close to one of their bi-annual releases, namely 3.0.25. No less than three security issues were unveiled. Two of them (CVE-2007-2446 and CVE-2007-2447) affect all currently supported Debian releases, namely sarge, etch and sid. One (CVE-2007-2444) affects both etch and unstable. This was the beginning of long days of rehearsal, helped out by Noah Meyerhans from the security team. Finally, updated versions for sarge and etch were uploaded to oldstable-security and stable-security for the autobuilders to catch up (they still are catching up because I uploaded the updates without the .orig.tar.gz file, which unveiled a bug in the security autobuilders apparently). These updates for sarge and etch should be available as soon as the security team completes the final checks and approval, I guess. And, today, the Samba Team gave us early access to the newly released 3.0.25 version of samba so that we could build and upload that package 2 hours before it was officially announced..:-) Thanks a lot again to the Samba Team for their care and more particularly to Gerald "Jerry" Carter (someone I have deep respect for, along with Jeremy Allison and Andrew Tridgell). Thanks to Steve Langasek for his hard work preparing the 3.0.25 release (tracking down an upstream bug in the Pythin bindings build). And, finally, thanks to the Debian security team and more particularly to Noah Meyerhans for his work on backporting some of the patches. This was a tiring but really motivating week-end. I hope that all Debianers who use Samba will enjoy the result.

AMD has announced that is going to deliver open graphics driver for ATI graphics card. Good news.

Sam commented:

"Some excellent questions, I might say. The 10% profit target that you mention is probably the target for Return on Capital Employed (ROCE), which was set at 10% by the Co-op Commission. This number is not displayed on page 13 of the annual review, but I think it is calcuable from the information there."

I also noticed that Sam's excellent site has a report on the Central and Eastern AGM.

Kevin Mark commented:

"Hear hear! I was waiting to read about the Eurovision results! But I'm sure AJ and Martin want a let's take the fun out of Debian planet debian tech and I'm sure someone can make them a custom feed for the cost of a reasonable bounty!

Nice to know I'm not alone. (Why mention AJ?) Joachim Breitner wrote:

"I once tried to start a German planet Debian because the official only allows English postings. Maybe it would be nice to have some kind of tagging standard that works across aggregators. This way, we could allow any kind of posts on p.d.o, and allow various sets to chose from, based on tags. I, for example, would choose (german OR english), others might chose (english AND tech), or ((french OR english) AND debian). Some (NOT politics). And some just (jokes) :-)"

Why didn't German planet debian work? I think there's a sort-of standard in dc:subject and xml:lang, but we'd need to agree a tag dictionary. And I think planetplanet generates files, so we'd need some cached or database solution. Update: Joachim Breitner replied:

"To be correct, I tried setting up an international planet debian on gluck with no language restrictions, but interest was rather low and planet (the software) is not really maintenance-free, so I stopped after a while. A German planet was tried by someone around Christmas, I think, but AFAIR he lost interest while trying to make it look pretty. I guess I should follow up on that. :-) (PS: Can you make the comment form store the information in a cookie? Would save some typing.)"

Sorry, these pages are static html at the moment. You could try using Saved Forms in Iceweasel/Firefox, which may be configured at Edit: Preferences: Privacy: Saved Forms or similar.

Sound Juicer "From a Council Flat Paying Higher Rate Tax" 2.19.1 is out. Tarballs are available on burtonini.com, or from the GNOME FTP servers. Brown paper bag release to fix not one but two bugs causing SJ to not start.

• Fix type names, which caused an assertion on startup
• Fix profile checking, which caused SJ to fail to parse any profiles
• Fix error handling when using errno (#438122)

I am currently with Optus, and while they were acceptable a few years ago, they have descended in the last couple of years so that they're chair-throwingly obnoxious now. So, as I really, really dislike giving retarded monkeys my money (I'd get rid of govco if I could, really), I'm looking to move. Specifically, I want a telephone communications provider:

• Whose phone monkeys are clueful enough to be able to execute the things I've requested be done, and can actually do what they tell me they're going to do (detailed story there is available on request);
• Who doesn't summarily raise line rental by 40%, and insult my intelligence by telling me that it's better because I get a pile of "included minutes" that I'll never use;
• That doesn't use one of those horrendous, useless voice recognition systems (they're slower and more error prone than just pressing the damn numbers) -- I'd love a real person on the line when I call, but I'm realistic enough to accept that that's just not going to happen;
• Who, if they do decide to outsource their phone support, at least sends it to a call centre who has people who don't have an overpoweringly strong accent;
• That just lists their charges (all of their charges) simply and easily, so I can say "that's what it's going to cost me to use these guys", instead of the endless twisty passages of bundling, "special deals" (which expire a couple of months after you sign up to them), and other customer-confusing effluvia;
• Who employ someone to proofread their website (I'm looking at you, Southern Cross, with your grocers' apostrophe

). Do I really have a hope in hell here? At this point, I'll pay a bit more for phone service if it doesn't involve getting screwed over at the telco's whim, and dealing with incomprehensible Indian call centres. Answers on a postcard, or (if you're feeling really technical) in an e-mail to mpalmer@hezmatt.org. We're talking about Australia here, in case anyone's wondering -- moving to another country for phone service is probably a bit too drastic. Switching to pure VoIP would be an option if I didn't need to have a fixed-line service for my ADSL.

Just after having made the 1.1.0 release I discovered an annoying bug in that version that meant using caption files made it fail in new and interesting ways, I've now fixed the bug and released bpgallery 1.1.1 which fixes this.

Yes, Microsoft have started swinging their patent bat a little harder. However, before everyone gets all screamy about the "235 patents" thing, read a little closer:

the Linux kernel [...] violates 42 Microsoft patents. The Linux graphical user interfaces - essentially, the way design elements like menus and toolbars are set up - run afoul of another 65, he claims. The Open Office suite of programs, which is analogous to Microsoft Office, infringes 45 more. E-mail programs infringe 15, while other assorted FOSS programs allegedly transgress 68.

While I can forgive a journalist the sin of not knowing the difference between Linux-the-kernel and Linux-the-entire-FOSS-ecosystem, can people who should have a clue about this please stop perpetuating the inaccuracy? Yes, 42 patents is still pretty bad, but it's not 235, and getting facts straight is what we do on this side of the debate -- the frothing and bullshit is what the other guys do. Once you start looking at the various categories of patent, it really does make it obvious that there's some straw-grasping going on here. If GUI elements really infringe on 65 Microsoft patents, it seems reasonable to assume that KDE and Gnome have a subset each, so presumably nobody gets the full brunt of that lot. Also, those 68 "assorted FOSS programs" patents probably includes such widely-used classics as this pearler of a program (chosen at random from Freshmeat for it's likely small userbase; no other ridicule is intended or implied). What galls me, though, isn't the bandying around of random inaccurate numbers, it's classy things like this:

[some Microsoft sock-puppet] refuses to identify specific patents or explain how they're being infringed, lest FOSS advocates start filing challenges to them.

Yeah, we'd hate it if the patent system were at all used in the manner in which it were intended. I don't suppose anyone else is getting SCO-style "we've got 100,000 lines of infringing code, but we're not going to tell you which ones" flashback at this point? Personally, I think that if Microsoft ever does starts with a patent smackdown, it won't be 235 patents, it'll be two or three, the strongest ones they can find, and 90 minutes after the plaintiff(s) is/are served the infringing code will be history, and we'll go back to living our lives as normal. That's even assuming that the patents are legit, which is by no means a given (even with the US' insanely broad patent provisions). As the linked article explains, the US Supreme Court has started getting a bit noisy on the subject of over-broad and pointless software patents, so perhaps Microsoft is just getting it's licks in before the Supremes take it's noisy rattle away from it. There's some side-commentary on this issue at Groklaw, and the Patent Law blog has a quick mention of it too (with all sorts of nice frothy comments, to boot).

Sound Juicer "Nikki's Growing A Patch Out In The Backyard" 2.19.0 is out. Tarballs are available on burtonini.com, or from the GNOME FTP servers. This is the first release in the 2.19.x development series, after I failed to do anything useful in 2.17.x...

• Attempt to repair brain-dead FreeDB encoding (Ka-Hing Cheung)
• Fix memory leaks in musicbrainz object (Cristian Persch)
• Follow the preferred sound device (Bastien Nocera)
• Use xdg-user-dirs (BN)
• Use gst_element_make_from_uri instead of hardcoding cdparanoia (James Livingstone)
• Display a better message after extracting (Adam Petaccia)
• Fix detection of FreeDB albums (David Mandelberg)
• Accessibility fixes (Rich Burridge, Patrick Wade)
• Don't crash if libnautilus-burn doens't know of a device (Pascal Terjan)
• Initialise threading earlier
• Convert strerror() output to UTF-8 for display (PT)
• Remove "Application" category from desktop file
• Fix srcdir!=builddir builds (Loic Minier)
• Internal refactor of extractor and metadata objects

I've taken over upstream development of xmms-scrobbler, it's now hosted at http://xmms-scrobbler.sommitrealweird.co.uk/ and has a public readonly git repository for the code. Since taking over development I've dropped a lot of the tag parsing code and replaced it with taglib, found some "interesting" locking issues and am now looking at making it so that we don't need to restart it when adding username/password. It'd also be good to get it to actually report a username/password breakage at least once rather than being totally silent about it! In other news, I've just released a new version of bpgallery, this version adds in medium sized pages - i.e. you can have the index page, the index page links to pages that have a resized image on the page, and has next/previous links. If you also generate full sized pages then the link from the image on the medium sized page goes to the full sized page, otherwise it'll go straight to the origional image. Also thinking about making a last.fm "client" for xmms so that I can keep all my music playing in the one place (lastfmproxy appears to have stopped working for me :/)

Had to create a new SPI CA as the old isnt usable anymore. You can find it here and the information to validate it is also available. You should include this in your browser by downloading the file, then checking its fingerprints against those found in the second link above, with those commands there and finally simply “opening” the downloaded file in your browser. Depending on your browser you will see a dialog with wildly differing details. You should be able to import it there and/or also assign a “Trust it for web/mail/whatever” level. Its currently used to sign certificates for various SPI sites, all of DebConf.org and also various debian.org sites (certs send out, may need a day to get installed everywhere) like alioth, db, nm, rt.

It's voting time in the lunacy that is the Eurovision Song Contest. Here is a summary, including some strange bits of Euronglish:

1. Bosnia+Herzegovina - nice but no chance;
2. Spain - 4 guys in tight white and pink lights - creepy;
3. Belarus - Medallion man with a rock ballad to bukkake - "when you cast your loving potion over me";
4. Ireland - Folksie, charming, dark yet jolly;
5. Finland - more noisy dark rock with drums;
6. Republic of Macedonia - up/down ballad with drums;
7. Slovenia - Opera in black with drums;
8. Hungary - Bluesy rock angst. No drums. Phew;
9. Lithuania - Guitarry song with "lonely shades walking";
10. Greece - an odd disco/dance at the "heart of attention";
11. Georgia - a nice slice of trance;
12. Sweden - revenge of 70s glam! Maybe "with my head in the can" was deliberate;
13. France - Franglais pop. How did monoglots follow this?
14. Latvia - a singalong. In Italian. But it looks like Italy aren't here this year. So how's that help?;
15. Russia - another fun girl band, singing different (ruder) lyrics to the official song sheet;
16. Germany - sexism swings;
17. Serbia - a red up/down song;
18. Ukraine - dance nonsense - the first song I walked out on this year;
19. UK - an example of Britishness and why I call myself English, not British;
20. Romania - a fun little polyglot guitar song;
21. Bulgaria - drums, dance, squeaking;
22. Turkey - "How much I crave for you" - a dancey whirly song;
23. Armenia - the tearful stalker's song;
24. Moldova - gothest song of the night - "each people will gnaw our wishes no more"???

I can't pick a winner from that. I only voted to try to stop Ukraine.

Apparantly not everyone knows binNMUs. Sometimes bugs are filed when binNMUs can solve the problem. Some people even wonder about breaking NMUs without any notice. A binNMU:
* no source changes * sheduled on the build daemons by someone with wanna-build access * for versions of packages which already exist in the archive * can be requested on debian-release@lists.debian.org, no bugs should be filed So a binNMU is used often for library transitions with ABI changes but without API changes (libFOO-dev has not been renamed, but libFOObar has been renamed to libFOObaz) or to fix an upload of a package that was not built in a clean environment. Note that it is a good idea to contact the maintainers of packages that reverse depend on your library before starting a library transition in any case so people know why their packages don't build/install anymore and maybe can even coordinate uploads they were going to make anyway. Informing the Release Team about library transitions of more than a couple of packages would also be appreciated....

Paul Graham makes some interesting observations about taking risks to achieve career benefits. One thing he doesn’t mention is that the risks have to match your life situation. If you are 21, living with your parents, and single (typical for a CS graduate) then you should take the riskiest options in terms of your career (apart from working in Iraq of course). If you don’t have much money then you don’t have much to lose. If you live with your parents then you still have accommodation and food even if you have no money. If you have no dependents (SO or children) then there’s nothing compelling you to earn a certain income. When you get older you may get a mortgage, a SO, and/or children. Also you won’t live with your parents forever. Most career risks that you might want to take aren’t possible if you leave them too late. Finally if you do something risky such as starting your own company and it doesn’t work out then it’s still going to look good on your CV. If you already have a lot of experience in the industry then the CV improvement may not be worth the time and effort invested in an unsuccessful company. When I was 22 I (along with two business partners) started an Internet cafe. It went reasonably well (by the standards of small businesses), it lasted for a few years before cheap net access at home killed most of the business. At the time the cafe had to close the ISP side of the business was doing reasonably well and one of my partners bought the operating ISP business. This buy-out caused me to approximately break even out of the entire business which is a lot better than most small businesses do. When I was 26 I moved to London (I have dual nationality, UK and Australian). The experience I had gained from running my own business allowed me to immediately get contract work for large ISPs in Europe. Most of the risks in my career were ones that I took while living with my parents. At the time I didn’t think through the issues of mortgages etc, my thinking was mostly along the lines of “it could work, I’m bored, so why not?”. ;) Update: While in the process of writing this blog post I forwarded the URL of a dating service for scientists (sciconnect.com) to some friends. The main page has pictures of single people wearing lab coats and using laptops which I found amusing. I have no idea whether it’s a good service or not, but the pictures on the main page made it worth a look. It seems that I accidentally pasted the wrong URL into my blog post so people who were looking for the Paul Graham article ended up at the dating service instead. But I guess if you are the type of person who reads my blog and who is interested in a link to Paul Graham’s blog and you happen to be single then a dating service for scientists might be of some interest. Thanks to MJ Ray for pointing out my error.

Do you know this?

you write a brilliant entry for a blog, a blog-comment or something in a textarea-field in your browser.

after finished, you press the Send-Button, and ... the browser crashes, or nothing happens because Javacript or Cookies were disabled in your browser.

the result is, the brilliant text you just wrote by investing time has gone to bit-heaven.

Has anyone a solution for this, beside writing in a text-editor, pasting it over to the textarea, fighting against linebreaks and things?

I once tried Editus Externus a Firefox extension, but that one produces a lot of problems in the fight iso-8859-x vs. utf8, and the linewrap problem is also there.

So how can i make sure that the things i type in a textarea aren't lost, when my cookie expires, Javascript sucks or my Browser dies?