Hello, I've noticed that Calculatin report stores the number of words translated in a segment, for which a Machine segment and a 40% fuzzy segment were suggested, inside the >90% category.
It looks to me as a bug, the numeber of translated word should fall in the Machine Matches category o in the Manually Translated one.
Thanks in advance,
Ok Gerhard, I'll contact my IBM representative to see.
Thanks for your time.
Hello Gerhard, thanks for your prompt response. I'm sorry but I don't think I can send you my folder because it is under IBM copyright. Is there some other way we can reproduce the problem, Can I contact you via skype, this way I could share with you my desktop?
Anyway I did further testing on my colleague's machine and was able to reproduce the problem. After that I removed the Machine Translation memory from the properties of the folder, and I noticed that a 39% fuzzy segment fell under the Manually Translated category. So it seems like the problem happens when OTM suggests both a Machine Match and a Fuzzy Match with less than 50% of fuzzyness.
Please let me know if there is another way for you to reproduce the problem.
when the folder is under IBM coyright, it seems you are working on an official IBM project. Why don't you open a problem report against OpenTM2 on the IBM internal problem data base. Or just send a mail to the IBM TMSupport eMail address. It will be no problem to append your sample folder to the problem report on the internal problem database.
you're right, the segment should be counted in the Machine Matches category.
This does not seem to be a general bug - in a self build sample such a segment was correctly counted in the Machine Match category.
For the investigation of this issue I will need some sample data. Could you please provide a sample folder containing the document of the segment and the memory being used for translation (it should contain at least the fuzzy proposal and the machine match for the segment)? Thanks!
OpenTM2 project community infrastructure is provided by beo Gesellschaft für Sprachen & Technologie mbH