Day 2 at Open Translation Tools started bright and early. And just as interactive as day one. Now that we all had a bit more knowledge under our belts, it made sense to go around the room again and ask each participant to identify something specific they definitely wanted to get to before the end of the next day. For me that was easy. I wanted to at least have set up an environment on my machine where I could potentially participate in localization of source code.

If you've never done that kind of thing previously you wouldn't even know where to begin. But now (or at least by the end of day 2) I understand that for most FOSS projects what you are really doing is taking a .pot file - effectively the list of all the strings of text in the program's interface - and translating each of those strings into the target language. The completed set of translated strings is saved as a .po file. So, one .pot file, many .po files, one for each localization. There are other systems, of course, but this describes the norm in FOSS projects.

Gentle reader, I'm delighted to say that by the end of day 2, I had indeed installed a .po editor (I used poEdit), downloaded a .pot file from a FOSS project and tried at least a few translations of some of the text strings in order to create a localized .po file. I learn in very small steps :-)

Once again this day we broke into parallel discussion groups. The choices were hard, since I have so much to learn, but for the first one I went with the group discussing challenges and opportunities in managing a volunteer translation community. There, I was inspired by Danilo Segan's account of managing the 300+ volunteers contributing to the Serbian localization of Gnome (one of the familiar desktop environments for those using a Linux distribution). In the same vein, Dimitris Glezos explained how the Greek localization of Fedora (the Linux distribution, not the institutional repository) works. At least on the surface, it was clear that the localization of software seems to be organized differently than the volunteer communities translating open content manuals or other text.

The second parallel session I participated in was focused on quality assurance, or quality assessment. Here the differences between localization efforts for software interfaces versus translation of free text again came to the surface. Professional translation of text always involves a careful layer of proofreading and editing. Our discussion concentrated on how to get at least some of that built into the workflow for community-led translation efforts. I doubt we fully cracked this problem, but at least everyone involved got a much sharper view of where the challenges reside.

The first SpeedGeeking session of the OTT07 took place immediately after lunch. Since very few will ever have experienced SpeedGeeking, I'll describe it briefly. Take 7 people with fabulously interesting projects or software tools. Give them each a table and whatever communications tools they desire. Divide the remaining participants between each of the tables (about 3 or 4 people per table). Then begin. The SpeedGeeker gets 4 minutes to describe his or her project. When time is called everyone (except the SpeedGeekers) moves to the next table. Repeat. In about 35 minutes you have been introduced to 7 exciting new projects. It is exhausting, especially for the presenters, but you will certainly have identified one or two people you absolutely must follow-up with later. Very efficient, and lots of fun.

I'm holding back here from describing each of the projects lest this post become impossibly long. Instead I'll do an individual post on each of the projects that impressed me later.

After a short break we went back to the parallel session format. The one I attended concentrated on workflow for translation in an open content project. What was clear was that most workflows were ad hoc, created through trial and error rather than on the tried and tested model of the professional translation industry. But of course the conditions are different.

The second day concluded with another communal meal and lively conversations that extended far into the night (and, for a few participant, in to wee hours of the morning).