Hi.

1. XLF files come presegmented by the tool, which created those. So to solve problem 1, contact the customer and ask for the kind of segmentation rules. This is the only decent way to change it. You can however also use the "Split segment" function - but I'm afraid your customer will moan, as they have most probably segmented that way by purpose.

2. Well, in my case there are also "seg-source" tags in the WorldServer files, but these are already present in the files I get from the customer. And they can import them with no problems...

Can i send you a sample of the file for your reference in order to try to import it to SDL Studio and see maybe this can be solved?

I just created a new XML filetype, imported the corresponding XLF, set "source" as not translatable and used the HTML5 embedded content processor. As file identifier you must set *.xliff instead of *.xml - than it works.

I will try that later on today and will let you know...

One other question - it's kind of a tricky one.

If we have a TM from previous work which contains these large segments with their translation, and now we are adjusting the segmentation - can this be done also on the TM? It will be pity to lose the previous translation just because we fixed the segmentation.

I am afraid that changing the TM will be not possible, at least not without editing it externally in TMX for example. But even there it will be tricky. I could imagine exporting TMX to Excel, than splitting the segments there by a macro (or via Word). In that respect I would not do that myself.

OTOH the translator can always use a lower minimum match setting and go for concordance.

Hi Jerzy,

I still have issues with setting the filetype.

Can you please provide a detailed explanation of exactly how to do that?

Thanks

And one more thing, will you suggestion of a new filetype will fix the segmentations?

Hi Jerzy,

I'm trying to follow your instructions as I need to do the same thing: process a set of Wordpress XLIFFs with CDATA sections in them. However, for some reason I can't see your screenshots above.

I created the new filetype (I chose XML-Embedded Content) but can't get it to recognize the file. What could I be missing? After creating the filetype I didn't change any settings (just clicked OK through the end), but added the HTML5 embedded content processor under "Embedded content".

Thanks!

Nora

Dear Jerzy and Paul,

I am working with XLF files and have the same issue as Fouad (i.e. The problem is that there are some very large segments, "<source> --very large segment--</source>" which includes few paragraphs. When I try to create a project in SDL Studio 2014, each very large segment appears as a single segment in the Editor. It doesn't do the proper segmentation as in a regular Word Document or any other documents.). Unfortunately, the screenshots you posted back 2 years ago do no longer display. I cannot find any suitable solution to this problem. Could you advise on how to help segment the files properly in Studio?

Thank you very much in advance

Géraldine (working in Nancy Matis team)

Hi Fouad,

You only have two elements to worry about in here, <source> and <target>.  Just copy the source to the target element in the xliff file itself using regex, and then create a simple xml file with one rule to extract the text from the target element.  That's it.  I have explained this in the past in detail in this article:

And if you need more information on how to deal with xml filetypes please refer to this one:

Between these two you have enough information to tackle this problem, especially the first one as it's exactly what you need to achieve.

Regards

Paul

Hi Fouad,

You only have two elements to worry about in here, <source> and <target>.  Just copy the source to the target element in the xliff file itself using regex, and then create a simple xml file with one rule to extract the text from the target element.  That's it.  I have explained this in the past in detail in this article:

And if you need more information on how to deal with xml filetypes please refer to this one:

Between these two you have enough information to tackle this problem, especially the first one as it's exactly what you need to achieve.

Regards

Paul