MetaBrainz Blog

Its been a busy summer behind the scenes at MusicBrainz! All this week I want to post a series of blog entries to catch you up on what we’ve been up to and how I expect us to move ahead with development of new MusicBrainz features. Over the summer we had three students working on various Google Summer of Code projects, making all sorts of interesting progress. In today’s blog post I’ll cover Oliver Charles’ (Acid2) work on his Template Toolkit branch. Later this week I will cover other SoC projects and current work on NGS in an effort to lay out a roadmap for future MusicBrainz development.

For lots of details on Oliver’s work, please check out his blog. Without delving into details here, I’ll say that his work is an important stepping stone in getting to the fabled NGS. Many people have commented that the current mb_server codebase is too confusing and too hard to setup. This has the effect of keeping developers away from hacking on the server. The general consensus was that we needed to clean up the codebase to make it easier to follow, separate business logic from the user interface logic and to use some more modern tools.

Oliver worked to address all of these points over the summer. We were clear on the fact that this project was too much to finish in the time that was available, so that the project would remain unfinished at the end of summer. Fortunately, MetaBrainz has been signing up more customers, so we finally have enough money in the bank to hire a coder to keep working on new server features. To that end, we’ve just hired Oliver part time to continue working on his project! Currently, the goal is to finish his Template Toolkit work by the end of the year and to start a beta test phase at the beginning of next year.

Hiring a developer to continue hacking on new features for the server is really important. My time in the last year has been taken up by keeping MusicBrainz running and keeping our search servers from melting. Unfortunately we’re coming up on a year without a server update, which is tragic. Having money in the bank should allow us to finally speed up the pace at which we’re updating our servers. Ideally I would like to have 4 updates per year — we may not reach that just yet, but I can see three releases happening in 2009!

Tomorrow I will post an update on Niklas’ collection work and the upcoming server release. Stay tuned.

For now, congratulations to Oliver Charles and welcome as an independent contractor to the MetaBrainz Foundation!

The BBC has just released its next MusicBrainz enabled feature: Dynamic artist pages.

You can see how often your favorite artists have been played on the BBC networks since last year. Turns out Coldplay is quite popular. To see how often a specific artist has been played, find the MBID on MusicBrainz and then go to this URL:

http://www.bbc.co.uk/music/artists/<artist mbid>

For instance, Portishead has mbid 8f6bd1e4-fbe1-4f50-aa9b-94c450ec0f11, so to check on Portishead, you’d go to:

http://www.bbc.co.uk/music/artists/8f6bd1e4-fbe1-4f50-aa9b-94c450ec0f11

The folks at the BBC assure me that all the artist MBIDs have a page there, but the page isn’t guaranteed to have much data. Even the MusicBrainz test artist is there. 🙂

Well done BBC! I’ve learned lots and lots about how the BBC operates and to see my friends at the BBC make highly visible progress tickles me pink!

Matthew Shorter of the BBC offers a little more insight:

Why are we doing this?
Currently our offerings around individual artists tend to be dispersed and hard to find. This leads to poor search performance for BBC music content, which means that users will typically only find content if directed by broadcast, or serendipitously by browsing brand sites.

Persistent unique URLs for artist pages which automatically aggregate what the BBC has to offer around individual artists will lead over time to much improved search performance and facilitate wider syndication of our content, building reach to brands. Automation and dynamic publishing means the pages can be created and maintained with a fraction of the manpower and server
load of the current generation.

Building good interrelated metadata for artists and programmes will also help greatly to enrich the music offering of radio & TV sites, offering such things as a chart of artists most played by a network, further information behind tracklists, rich now-playing information and so on.

What’s the offering?
We have a page for every artist in the MusicBrainz database – c.350,000 and counting. They contain, where available, basic information about the artist, discographies, high-quality images and details of play count by BBC networks & programmes. (It’s worth pointing out that for most of these pages, most of the time, there won’t be much content, but that’s fine, because pages will only ever be linked where we have broadcast or otherwise featured an artist, which by definition makes them significant.) See instructions at the bottom of this mail for how to access a given artist page*