Start a new topic

QA features request: segment size and glossary

Hi. I have noticed that, personally, my most common error while translating are omissions.


For that reason, I would suggest implementing a few new features in QA: the first is a quick check to compare source and target size (as in a percentage: 25%, 50%, 100% etc). This would alert the user when they might have forgotten to insert text in target.


The second is to check for the presence of specific words in target, i.e. if the user defines a glossary entry and that entry appears in source, then the corresponding term should be in target. This would really me avoid terminology omission errors.


Lastly, the bookmark segment function is pretty cool, but how do I remove the "B" highlight once I'm done checking that segment? Thanks!


Luisa: ...my most common error while translating are omissions.


If you use insert Auto-Assemble - CafeTran's core feature - omissions are virtually impossible.


This would really me avoid terminology omission errors.


An omission in this sentence maybe... But again, insert Auto-Assembly would avoid this. I strongly suggest to avoid "glossaries" though. Use TM for fragments instead.


H.

> quick check to compare source and target size


In the meantime, while waiting for this, try CheckMate that has a pretty nice GUi for that (concerning features, this appears quite useful).


image


And yes, it would be nice to stop CT confirming empty segments (a thing you usually hardly need) without further ado. These can be found by QA > Empty segments, but this is not the best way to avoid this kind of accidents.


> presence of specific words in target


Menu QA > Term consisteny check (glossary)

wooden, how do you perfect the use of the Auto-assemble function? Unless I have a highly similar project, it's usually just gibberish (a mix of English and Portuguese terms and Portuguese [target] terms are never in the right place because the grammar order is the opposite of English). Besides, it takes some time to load, time I could use to write my own translation. So, I haven't really found it to be of much use, but I am curious to hear your ideas.


Igor, regarding terminology, how does it work? When I add terms to "non-translatable fragments", does the terms consistency function check to see if those are present in target as well (for instance, brand names). And does it do reverse check, that is, start from a term in the target and check to see if it's in source? How can I optimise the elimination of omission errors, much the same as I already have for punctuation, tags, etc?


Thanks to tre as well for CheckMate. Although it would be cool not to have to use an external app to do QA checks for length consistency between source/target.

Luisa: how do you perfect the use of the Auto-assemble function?


By adding as many words/fragments to your relevant resources as you can. It won't work overnight, plus:


 the grammar order is the opposite of English


Nothing much you can do about that (except for adding those fragments, of course). But at least everything will be there, either as the correct Portugese term (settings!), or as the untranslated English term, and since your question is about "omissions", that should do. Like me (mainly ENG>DUT), you may have to typ everything from scratch (the assembled translation deletes itself the moment you start typing), but at least you saw the correct terms in the AA, and you are less likely to omit something.


it takes some time to load


Well, it shouldn't take long, so maybe you should have a look at your settings and features (RAM, Recall Memory).


Trouble is, I don't think your proposed solution - compare sizes/number of words/characters - will work. It will be a different paramater for the various language combinations,  and I have no doubt you'll end up with wagonloads of false positives.


H.


> check to compare source and target size (as in a percentage: 25%, 50%, 100% etc)


That's a good idea to add this QA check.


> The second is to check for the presence of specific words in target,  i.e. if the user defines a glossary entry and that entry appears in  source, then the corresponding term should be in target. This would  really me avoid terminology omission errors.


See QA > Terms consistency check (Glossary).


> but how do I remove the "B" highlight once I'm done checking that segment?


Right click at the source segment editor and choose the Bookmark segment again  to toggle the bookmark or use the keyboard shortcut for toggling bookmarks.

> "non-translatable fragments"


I plan to add this QA check soon too. Thanks.

It would be nice to have the "Warn if segment longer than" + lenght in chars implemented, too (see Checkmate screenshot above). This is one of the most common scenarios for localization.

Login to post a comment