tfennell

How you deal with this may depend on the file format. Do you have a sample... just a small one with a couple of paragraphs?

Hey Paul!

I am just talking about plain ole' Word here.

Search for the word фактическими and you will see where the segment breaks before the word данными.

There is only a page break here, no paragraph markers.  It would be much better to keep the whole sentence as the segment, because on other pages or in other documents, the sentence may be in the middle of the page, not split over a page break.

BTW last night I also replied to a comment you made several years ago about horizontal (source above target) viewing in editor.  I'll try to see if you replied, but I'm not sure how to find it now to see if you replied.

Did you see it and the link I included?

DOCX

Oh, I did find it - it is in the thread

RE: Top-Down layout in Editor View using a plug-in?

Looks like you did not see it yet - take a peek - I included a very interesting link :)

:

BTW did you see the stunning photograph in the New York Times link in the section "Who Get's a Loan?  Who Doesn't?"

https://www.nytimes.com/2020/04/26/business/ppp-small-business-loans.html

Sorry for giving you a huge file.

Attaching here a smaller example.

DOCX

tfennell said:

BTW did you see the stunning photograph in the New York Times link in the section "Who Get's a Loan?  Who Doesn't?"

Very cool!

tfennell

Thanks for the small file.  I don't think you have a solution in Studio though because these page breaks are not simple character codes, they are actual tags and Studio will move these into an external position and split the segment either side.  If you view "All Content" you can see them:

The best I can suggest is you search replace them in Word:

So search for ^m and then replace with nothing at all.  That will remove all the page breaks in your document and if you wanted Studio to ignore them anyway this may be the way forward. The only proble being putting them back afterwards.  If you need to do that perhaps replacing ^m with something like (XXX) for example would help as you could repeat the process in the target but replace (XXX) with ^m. Not great for your TM but

Quite a tricky one I think... how did the document get produced like this in the first place?

Dear Paul,

I see now how you are baffled by something that is absolutely normal for our workflow.

The matter is that we very often convert scanned pdfs into Word.  In the original document in Word, there was no page break and the sentence/paragraph flowed over into the next page, but we put a page break in so that we can keep parallel pagination in the final source and target files.

I think the answer may lie in replacing ^m with XXX or probably better ~~~ as an automated pre-edit in Word (we already have a macro that does a large number of pre-edits to prepare a file for optimal MT) before it goes into Studio.  The key thing is manually editing (or having a macro) after export back from Studio into Word to replace the ~~~ with page breaks so that we get our nice parallel source and target pagination again in the target Word file.

This will still be much easier than just putting page breaks in the target file by comparing it to the source file after translation, which can be very tedious, especially at the end of a lengthy translation.

In Studio we will get a slightly messy fuzzy match, but it will be clear what it is with the ~~~ and better than splitting the sentence into segments.

A bit of a messy workaround, but we can standardize this and make it workable.

My long term question would if Studio doesn't break on other tags, such as <bold>, why are other tags absolute segment-breakers?  I read your other posts on line-breaks, and it seems these can go either way in Studio.  Are paragraph markers more like page breaks or line breaks?

Thanks for taking a look at the NYT article - that was fun, and I felt it was important to speak up and share some of the good news in this crisis

Head's up - I had a huge nightmare with Language Resource Template last night, but I will be posting that in an additional thread.

Thanks again for your excellent advice and insightful question.

Best regards,

Tom

Actually, we are manually inserting the page break, so we can just manually insert the custom page-break character symbol(s) instead of putting in a page break at all.  We well still need to replace it in the target translation after export from Studio, but once again, this is easy.

A further reflection - this applies not only to scanned pdfs, but whenever you want to keep target file pagination parallel to source file pagination - even if this is a Word file.

Much easier to mark this in the source rather than try to insert page breaks after the target file has been produced.

A further reflection - this applies not only to scanned pdfs, but whenever you want to keep target file pagination parallel to source file pagination - even if this is a Word file.

Much easier to mark this in the source rather than try to insert page breaks after the target file has been produced.