Create a project from plain SDLXLIFF files?

Hi,

Some clients don't send packages but plain SDLXLIFF files for translation. When we get many of these individual SDLXLIFF files, we would like to group them into a single project.

So far I created a new project in Studio and specified the SDLXLIFFs as source files. This seem to work, but the files are altered slightly during project setup in Studio, which makes me a bit nervous. Therefore, I replaced the altered SDLXLIFFs with the original ones that I received from the client.

So I'm wondering what's the best way to create a project from a bunch of SDLXLIFF files?

Thanks
Frank

Parents Reply Children
  • Hi Evzen,

    I just did a quick test and the file did not get resegmented when a new TM is added with a different segmentation rule.

    Project 1 - Word file prepared with the Prepare without Project TM task

     

    Project 2 - SDLXLIFF file from Project 1 added to a new project with the Copy to Target Languages batch task. This project has a new TM that has an underscore segmentation rule.

     

    Just to check that the segmentation rule is working, I added a copy of the Word file to Project 2, and used Prepare without Project TM, with this result:

     

    So I think this shows that the SDLXLIFF was not resegmented when the Copy to Target Languages task is used.

  • Coincidentally, I learned about the re-segmentation the hard way :-( last week...
    The scenario was like this:

    • you have existing project with language pair A-B using TM with removed "Colon" segmentation rule
    • you add new target language C to the project (by editing Project Settings) without adding TM (with the 'special' segmentation rules setup) for that language right away at the same time, then click OK in the Project Settings
    • then Studio says "hey, I see you added new language" and offers preparing the files, i.e. basically running the "Convert to Translatable Format" and "Copy to Traget Lanagues" tasks
    • now, if you let Studio do it, it will use the DEFAULT language resources for the segmentation... obviously, since it has no other choice...
    • so you will end up with incorrectly segmented targets
    • so... you delete the bad target files
    • then you add the correct TM (for A-C language pair) to the project
    • and then you run manually the "Copy to Target Languages" batch task... (i.e. task which uses the source SDLXLIFFs, not the original source format)
    • ... et voila, the targets are now correctly segmented!

    EDIT:
    This was with MadCap Flare files, i.e. using XML file type (these weirdo files pretend to be HTML, but use some custom tags, so they are no more HTML-compatible and must be treated as generic XML, though they are like 99% HTML). It could be that other source formats behave diofferently. Who knows what's going on inside Studio...

  • Oh, I see. I followed your steps and got the same results with a plain Word file, even on a project with a single language pair, so it doesn't seem to be filetype-related. However, it seems to have this effect only in cases where the project was prepared from original source files, and not with SDLXLIFFs, that is, in my tests, if there are only SDLXLIFFs and no Word files in the project folders, the files don't get resegmented. I think for the purposes of the original question in this thread, this is an important point to keep in mind.