Elements of a Collaborative TM Environment

Last week I was in Québec City for the ATA-TCD Conference, which was superbly organized by Rina Ne'eman and Grant Hamilton.

The biggest takeaway of the event was the last presentation of the last day: A panel presentation by Don Shin, from 1-Stop Translation, and Rocío Txabarriaga, from Common Sense Advisory, named "The Future of the Translation Industry: MT, TM, Open Source, Crowdsourcing: Where’s It All Headed? And What Should You Do to Prepare?"

For his intervention, Don compiled some of the major efforts being done in those areas, but what I liked the most was his depiction of what the desktop of a translator working in a collaborative manner might look like.

The key point is that the translator is in control. At the top, you have the source text. Right below it, you have the translator's TM, the project's TM, and a Machine Translation of the segment. And below that, the translated segment.

It is up to the translator to choose which one of these sources she is going to use. On the right panel you have access to terminology and a chat window, to ask for help in live mode to other people working on the same project.

Finally, on the bottom right, there is a fare meter, that shows how much money the translator is making on the project. Whether this is a motivator or a demotivator depends on the price that the translator is getting.

Another panel discussing Translation Management Technologies, moderated by Duncan Shaw, failed to address what all the LSPs in the room were looking for: Interoperability. What I heard LSPs saying is that they want to let their translators work with any tool that they prefer (Trados, SDL, MemoQ, Across, whatever) and not to require them to have different tools for different projects.

What the industry seems to want, and the technology providers can't seem to be able to deliver, is a standard format for Translation Memories that does not get corrupted if you change from one tool to another. Like as Comma Delimited File that can be opened by Excel, Lotus, MySQL or Oracle, without any data loss.

I guess that was the original promise of TMX.