Translate
I am trying to use alignment files for PerfectMatch by extracting the sdlxliff from them. Are these sdlxliff files "normal" sdlxliff files in every respect?
Daniel
RWS Community
I am trying to use alignment files for PerfectMatch by extracting the sdlxliff from them. Are these sdlxliff files "normal" sdlxliff files in every respect?
Daniel
Are these sdlxliff files "normal" sdlxliff files in every respect?
I guess it depends what you mean by "normal". They are not associated with any particular filetype and this may be problematic when attempting to use for PerfectMatch. I haven't tested this to be fair, but I wouldn't expect it to work.
Paul Filkin | RWS
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
Are these sdlxliff files "normal" sdlxliff files in every respect?
I guess it depends what you mean by "normal". They are not associated with any particular filetype and this may be problematic when attempting to use for PerfectMatch. I haven't tested this to be fair, but I wouldn't expect it to work.
Paul Filkin | RWS
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
I can use them if I opened the alignment and confirmed the connections in Studio. I can't see how that could be automated (run from the command line or so). I can't go through 200 alignment files like that! If I can't automate this, then I can't use this approach.
I am still chewing on how to handle bilingual XML files, and I could easily auto-align all alignment files that only have 1-2 segments (the vast majority). I can read the number of segements from the SDLALIGN file. But all I could do with them then would be to make a TM.
Usecase is website exports (of course) where the translations may have happened in Studio, but where texts have been edited (SEO optimized etc.) in each language separately. Now a number of changes are done in the source language which have to be reflected in the target language without sweeping all optimization away that has happened in between. Same source segment text is very likely to have different translations, hence TM approach is a bit doubtful.
As far as I can see at this point, the alignment & perfect match approach will be good for long pieces of text which need to be segmented to make any sense. For shorter bits of text it will be the bilingual Excel file type, which improved a lot recently (CU6?).
How do other people do this? My dream is a bilingual XML file type that offers alignment if one segment contains text that would be split according to segmentation rules.
Daniel
I can use them if I opened the alignment and confirmed the connections in Studio.
Can you use the SDLXLFF Toolkit to confirm them all?
My dream is a bilingual XML file type that offers alignment if one segment contains text that would be split according to segmentation rules.
We have an item on our todo list to create a bilingual XML filetype seeing as the 3rd party developers who did this didn't want to share on the appstore. So when we come to this we'll consider your segmentation request. I'll add a link to this thread so we don't forget.
Our item, for your reference is this: SDLCOM-1459
Paul Filkin | RWS
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
Thanks for the hint with the SDLXLIFF toolkit - it worked! The next problem I encounter is matching these documents by folder.
By which criteria does this matching happen? Only by file name?
If I assign each file its counterpart individually, PerfectMatch works beautifully, but I have projects with hundreds of those files ...
We have an item on our todo list to create a bilingual XML filetype
That would be fantastic, especially since you have all the components of this workflow already, you just need to arrange them into the workflow!
From a user's point of view, I could imagine a file type that gives you the option to perform a segmentation at all or not. Then to give some parameters for a segmentation. One could say "automatically confirm if all alignments in a file are good (green line)". Or "automatically confirm all alignments if file has less than 2 segments". As always, the "devil is in the detail", as we say in German, and I am sure there would be problems to solve, but the feature would be very useful.
I can mail you my workflow if I get it to work. The last holdup - I think - is the matching by folder.
Daniel
I looked at both sdlxliffs of one source file: Once the sdlxliff file in the target folder, the other one the sdlxliff in the sdlalign zip-folder. The target folder sdlxliff (which Studio would expect in a PerfectMatch) differs vastly with much more context information, which I guess is used in the PerfectMatch process. In my case, the filename is indicative enough, but I think this approach hit an impasse here... Too bad, this would have been slick.
Daniel