Mayhem & Chaos Blog

 

Conferences

OSCON Blogging

All of last week I was in Portland for the OSCON conference and Tim O'Reilly invited my to guest blog the conference on the O'Reilly Radar:

• OSCON day 1: Beyond REST? Building Data Services with XMPP PubSub
• OSCON day 1: An Open Source Project Called "Failure:" Community Antipatterns to Know and Avoid
• OSCON day 2: Prophet, your path out of the cloud
• OSCON day 2: Do You Believe in the Users?
• OSCON day 3: Reflections on OSCON 2008

It was a nice change blogging on a high profile blog -- my blog posts had a much greater reach. Thanks Tim!

 

Copyright

Thoughts on Copyright Registries

The Creative Commons invited me to sit on the "Developers of digital copyright registries and similar animals" panel tomorrow at their Creative Commons Technology Summit. The CC is asking the question: Should we create a copyright registry so people can find CC licensed works? The summit tomorrow is designed to answer this question and I'd like to post some of my thoughts here.

When I first got involved with the CC shortly before their launch, I applauded their intent to create a copyright registry. After all, what good is a license when you can't find the licensed works? Unfortunately the possible liability issues in operating such a registry (e.g. if someone fakes a CC dedication the CC could get sued) killed that idea. Now, six years later, I think the CC should not create a copyright registry. Live and learn -- my experience with MusicBrainz has taught me a few things that make me believe that the CC hosting their own copyright registry is not the best idea.

One issue is with scope and attention -- operating a registry on the scale of MusicBrainz is a lot of work and costs a bit of money to host. Unless the CC can make this copyright registry sustain itself, it could be a significant money sink.

More importantly, consider this: Findability == good metadata.

Take Flickr for instance -- there are gazillions of pictures. Those pictures that have tags and descriptions can be found in a meaningful manner and those that don't can't be found. Of course there are tools that let computers analyze what the colors in an image are, but they can't really tell us that a picture contains a beach and seagulls. Content without metadata may as well not exist -- if you can't find it, its of no use to you. This, of course, is the crux of the problem.

But creating good metadata systems that collect useful information is really hard. Most content creators cannot be bothered to properly provide metadata for their content. Just saying the word meta will cause most people to glaze over their eyes and wish they were doing something else. And forcing people to enter metadata is a perfect recipe for gathering crappy and useless metadata. How do you collect good metadata then?

Get fans to collect and manicure the data! They're engaged in the works created by their favorite artists. If you can provide the fans with a value proposition that makes sense, they will go to great lengths to collect good metadata. MusicBrainz has a constantly improving metadatabase because fans can organize their music collection using tools driven by the music metadata. Fans manicure a little bit of data at a time and reap great benefits of a cleaned up music collection. This value proposition makes sense and when coupled with peer-review of the data, the data improves constantly.

This process is quite challenging to establish -- its not about the data or software that collects the data -- its about the community of people who are working to manicure this data. Without the community you can do nothing but collect rubbish. MusicBrainz has restricted itself to one knowledge domain -- the CC aims to cover all types of works. Fostering communities to cover all aspects of knowledge would be much harder than covering only music.

I think the CC should not create its own registry, but instead use existing databases and connector technologies like DBpedia. DBpedia has managed to connect a lot of databases already (including MusicBrainz). The CC should identify/create/extend standards, protocols and best practices for creating, operating or augmenting existing projects to include the aspects of copyright registration it would like to see supported.

I believe this would be the best use of the CC's efforts, since:

1. It leverages existing projects/databases
2. Pushes the bulk of the work onto other projects
3. Many disparate data sources exist already and tying those together increases to total worth. (see metcalfe's law)
4. Gives a meaningful purpose to the Web-of-data (formerly the Semantic Web)

This approach makes the most amount of sense to me -- but then again, I'm a big fan of bottom-up systems. I'm looking forward to the summit tomorrow!

UPDATE: Yves points to a better map of the Data Web

 

Dumbshit Dept.

Proof that media company's lobbying efforts have gone too far

The Canadians are thinking of taking the trend of border patrol insanity one step further. As if the security theater isn't enough to make people shy away from traveling, the normally level headed Canadians want to have border agents check your iPod for illegal content!

WHAT?? Seriously, they are proposing to inspect people's iPods for infringing content. I keep checking the date, but sadly its not April 1. I can't begin to fathom the problems associated with this:

• Massive travel delays -- How long is a border agent going to spend inspecting someone's 160GB iPod with 40k songs on it?
• Different countries -- What if you and your iPod come from a country where you can legally copy music off the internet and into your iPod? Isn't that/wasn't that legal in Canada?? Are border agents supposed to be in tune with all the IP laws from all of the countries?
• Border agents are unsuited for this task -- How the hell are you going to train the bonehead inspectors to even have passing chance at doing this job right? The things they are asking of border agents require not only technical knowledge (imagine the horror of these idiots having to use all sorts of music players, not just iPods!) but also IP knowledge form all around the world.
• Invasion of privacy -- Well, privacy has been lost a long time ago.

This is the most asinine proposal I've ever heard. This proposal is not based in any sort of reality -- can you imagine the damage it will cause to Canada's tourism? And the amount of tax money spent to enforce this? The money spent on this program and the damages from this program would dwarf the entire media industry in Canada. Save a broken incumbent industry by lining their pockets with tax dollars -- very clever for the media companies. Too bad it has far reaching and serious implications for the whole country.

Media company lobbyists need to be shot. Plain and simple -- seriously, who thinks up this shit? I'm confident this proposal won't go anywhere, but its really embarrassing for Canada that this proposal ever made it into the public light.

 

Dumbshit Dept.

Email troubles

The heat wave in California over this past weekend fried the disks in my community mail server and I lost all the email from over the weekend. If you sent me mail over the weekend, (either to rob [fat] eorbit [dork] net or rob [fat] musicbrainz [dork] org) please re-send it so I can respond to it.

Sorry for the hassle and thanks for your understanding.