Hello Takaaki Yamane,

It isn't possible to filter directly in Studio alphabetically/change the order of the segments, as Studio uses the source file in the memory- rather than a database.

Perhaps if you are using excel (as source) you can first prepare (filter) outside of Studio- translate the file in Studio (updating your memory and then finally pretranslate the original source file through your TM.

I see. I find this a major limitation. Is it possible to extract/filter the same source segments so that I can check for inconsistency for corresponding target segments?

Say, there are 5 same source segments, and you want to check if the 5 corresponding targets segments are exact same in terms of phrasing wording and  punctuation. 

You can filter in Trados Studio. There are several ways to filer, one of them is:

- Go to Review tab.

- In the Display Filter section, insert a text (In Source or In Target).

- Press Enter.

I filter quite often to check that share come text.

Also, you can press F8 (or click the Verify button in the Review tab) to run a verification test. In the Project Settings > Verification > QA Checker > Inconsistencies, you can adjust this particular QA,

Go to Review Tab, then select the Display filter "Repetitions - all". This will show you all repetitions in the document in the order of appearance.

The QA checker has also an option for checking of consistency, which does exactly what you want.

Alphabetically sorting is indeed not possible, but you can use many other options to achieve what you need. Personally, I do not see a big drawback (if any) in the missing possibility of sorting. This would lead to miss a content and even identical segments may need different translation due to different context.

Repetitions=All does not filter for repetitions.
It just non-filters the content (vs First occ. and Excluding first occ.).

Sorting in TM mode would really be a useful function.
Removing semi-automatically redundancies would be a useful function.

Alternatiovely: showing ONLY repetitions in Editor mode.

JF Delannoy 

JF Delannoy said:

Repetitions=All does not filter for repetitions.
It just non-filters the content (vs First occ. and Excluding first occ.).

Can you please explain exactly what do you mean? Does the filter not work at all?

Sorry -
it keeps repetitive segments, and in my current project, many are so it looks like the unfiltered text.

But...

one would like/need to see them grouped.

Maybe with Xbench.

It would be intuitive and natural to llst segments alphabetically (in Editor mode, and in TM mode)
and would not cost much engineering.

JF Delannoy 

Studio does not have any options for sorting segments, as it works file based. This brings numerous advantages like easy dealing with files, moving them wherever you want, editing however you want etc. If you want sorting, use memoQ. This program works with databases and has its own advantages. But it does not offer this kind of flexibility Studio does. So as always in life—you cannot have a cake and have eaten a cake. Either you have this advantages or those. But you cannot have all of them in one piece of software.

Where can I read technical (computer science) details about this file-based mechanism?

What aspect of the file-based orientation prevent data sorting here?

Wikipedia, for what it's worth, states
"When creating a new (file-based) translation memory, Trados Studio creates a database file in which all translation units are stored."

Maybe the DB is segment-level, but with no notion of character strings?

JF Delannoy 

The topic was to sort segments in editor. This is not possible, as the editor works file based.

Sorting of TM in Studio is not possible afaik. But you can sort a TMX alphabetically in Olifant with no problem: https://okapiframework.org/wiki/index.php/Olifant

I still fail to see this big advantage of sorting, though.

JF Delannoy 

The topic was to sort segments in editor. This is not possible, as the editor works file based.

Sorting of TM in Studio is not possible afaik. But you can sort a TMX alphabetically in Olifant with no problem: https://okapiframework.org/wiki/index.php/Olifant

I still fail to see this big advantage of sorting, though.

Linguistics-wise: glossaries are routinely alphabetically sorted.

This eases casual lookup and tasks of cleaning out discrepancies and doubles.

TMs could be viewed as a special form of terminology.




Are there computer-science details about the file-based implementation, for the technically curious?




NB
I am a linguist (i.e. doing R&D in semantics) and computer scientist (Symbolic AI for text analysis).,
and have been initially very susprised by the absence of any syntax or semantics in MT tools at large.

JF Delannoy 

JF Delannoy said:

Are there computer-science details about the file-based implementation, for the technically curious?

All you really need to know is that Studio reads the SDLXLIFF file directly.  It doesn't import it into a database where sorting would be trivial for example.  There are advantages and disadvantages to this approach, but that's what we have.  To be able to make sorting work we would either have to read the file into a database, which isn't a trivial amount of work now that we have the platform we do, or we have to load it into memory and allow you to sort etc. and write back any changes after they were made.  Also non-trivial.

The sorting of the data in a TM should absolutely be technically possible, worth creating an idea for this.  It would also be quite feasible for anyone to write their own application to handle an SDLTM in this way... it's SQLite and the current editing features are basic anyway.  I'd really like to see one!