I was privileged to have been invited to Biohackathon 2009 in Tokyo and Okinawa. Greg von Kuster and myself from the Galaxy team attended. It was a very interesting event, part hackathon and part unconference, with a lot of satellite discussions and informal talks. While I didn't end up doing as much hacking as I hoped, the opportunity to discuss some big challenge areas from a technical perspective was useful. The venue was incredible, and DBCLS/OIST were incredibly generous in hosting us for so long in such amazing accommodations.

The best surprise from the event was meeting with people from DBCLS and learning how extensively they are using Galaxy. While I knew Mitsuteru Nakao and his colleagues had been working on translating Galaxy, it was impressive to see how far they had come. One of our major accomplishments for the meeting was Greg integrating the i18n work into the main Galaxy. We hit a number of snags along the way, and learned more than we wanted to about i18n. In particular, we discovered that there are two very different issues that needed to be dealt with:

  • Localizing the interface: Galaxy will now respect the Accept-Language header, and if an appropriate translation is available, translate strings in templates. Work still needs to be done on making sure all template strings are translatable, however all the framework pieces are in place. This has already been deployed, and if you visit [http://main.g2.bx.psu.edu] with Japanese as your preferred language, the interface will be translated.

  • Dealing with multi-language data: Regardless of what interface language is being used, it should be possible to enter any language in descriptive fields (like dataset info), tool parameters, and datasets themselves. This required making every level of the application unicode aware. This was also completed at (and immediately after) the hackathon, and it is now possible to use any unicode character in info fields and datasets. Such character can also be provided in tool parameters, but they are not yet passed down to the command line. There is still some work to do on ensuring that these characters are encoded safely.

I'm extremely happy with the results so far, and it is fantastic to be opening Galaxy up to a wider audience. More importantly, it makes me very happy to have more groups contributing to the Galaxy framework. It was really great to interact with the DBCLS group, and and I look forward to continuing to work with them to make their database resources more widely available through Galaxy.

Other accomplishments: Greg and Todd Harris made a lot of progress on Galaxy / Wormbase integration. I (finally) for runtime parameterization of workflow working, as well as first steps toward supporting selenium tests (getting them working under buildbot is another question).

All-in-all, a wonderful event. Regardless of how much hacking was done, for us this was a great exercise in international communication. Thanks so much to the organizers!