Thanks for your post. That sounds like an interesting idea.
Have you tried just uploading the TMX itself. TM-Town will automatically extract terms from both the source and target of the TMX and rank those terms using NLP techniques. TM-Town's glossary builder will then place those terms in two columns (source on the left, target on the right) ordered from the highest ranking terms to the lowest. You can then drag and drop target terms to match the source term in the GUI, or you can just type in the corresponding target term. As the ranking of terms from the source and target tend to "generally" match up it can often be very easy to create a glossary as the target term will appear just to the right of the source term (please see this blog post for some screenshots: https://www.tm-town.com/blog/creating-multilingual-termbase-now-simple-stress-free-process).
However, maybe you have already tried the above and it was lacking hence why you suggested the list of source terms. If so, I can definitely explore your idea of sending along a list of source terms.
If possible let me know how extracting terms from a TMX by itself goes and if it is lacking let's explore the idea of sending along a list.
Hans CafeTran Wiki
It's nice to know that you are reading here too :). The integration of TM-Town into CafeTran is reaching new horizons.
I have this question, related to (quickly) building a glossary from a translated project or a TMX file.
It is quite easy to generate a list of all possible variants of a source term, via a regular expression in the Find/Replace dialogue box. This list then is nicely displayed in the tabbed pane:
Here is my question: Can this list and the corresponding project and/or TMX be uploaded to TM-Town and will your NLP techniques allow us to quickly match the corresponding target terms?
That would be an absolute killer feature.
Are there any costs related to this for us users?