Hey, I just e-mailed you. I've ended up starting to prototype libtm in Python -- getting started in C just wasn't working. I still "think" in Python. Python works out well because all the libraries I'm using (leveldb and ICU, mainly) also have very nice Python bindings. Once I've got the prototype in Python, I'll port it to C.
I'll see if I can put what I've got in Python up on github after I finish this big project at work this week and next. I'd be really interested in your input. Did you read the Tim Baldwin presentation on translation retrieval I referenced in my post on libtm? He's done some really cool research that has had a real impact on the way I keep my TMs (e.g., trying not to go over 20,000 segments).
I'll see if I can put what I've got in Python up on github after I finish this big project at work this week and next. I'd be really interested in your input. Did you read the Tim Baldwin presentation on translation retrieval I referenced in my post on libtm? He's done some really cool research that has had a real impact on the way I keep my TMs (e.g., trying not to go over 20,000 segments).