Interview: Kuno Woudt, MusicBrainz

In Amsterdam I met up with Kuno Woudt, better known to many of us as warp, developer for MusicBrainz. He is currently building the next generation of the MusicBrainz website, which supplies metadata on music and its provenance to music player software, companies, and organizations like the BBC.

How long have you been involved with MusicBrainz?

I've been a user for about 5 years — mostly entering music, tagging music, etc. — I've been a developer for about a year, but I officially started working in February 2010. Before I could get started, I had to prove my worth by trying to get some code checked in as part of the interview process.

Tell us a little about the existing site, and what free software you're using to build the new site?

I don't know too much about the existing site, other than it's written in Perl... The site was started about 10 years ago and web development, especially free software, has moved on hugely since then. The new site is written with Catalyst, which is a web framework for Perl, similar to other frameworks like Ruby on Rails or CakePHP. We also use PostgreSQL for our database — as much as we'd like to, we're unable to support other databases because we have some modules that are written especially for PostgreSQL, mostly for collating Unicode data — we have albums and music in languages like Japanese and Korean, and presenting those to an international audience is tricky, so we use ''libuca'' (Unicode collation algorthim), which is free software.

If people want to get involved with MB, what are things that they could do?

We have developers who are being paid to work on the site, but we can always use more community help. One of the main uses for MusicBrainz is people tagging their music using Picard (written in Python) — Picard is entirely written by volunteers in the community, and we can always use more help there. If you know a lot about music, or you have a lot of records and CDs, you can help us think more about how to model the world of music in our database... for example: a song featuring multiple artists may be released in several ways — figuring out how those releases are labelled on the various CDs and records is a huge help. Classical music is especially difficult because this is not just songs and albums, but a much larger body of work with a different structure, because much of the music is in the public domain — there is not just one release. We can use more help capturing that and storing it on our database.

All the software written by MusicBrainz is free software, and free software is used to run and develop the site. Download MusicBrainz software at their website, http://musicbrainz.org/.