Forum: MemoQ support
Topic: Free workflow for handling memoQ projects containing a lot of repetitions
Poster: Stanislav Okhvat
Post title: Comments about the workflow for translating projects with lots of repetitions
Hello, Stepan,
Thanks for all the comments.
I agree, the word "impossible" is not suitable here.
As Stepan correctly said, it is in fact possible to translate a project with a lot of repetitions using a view of First Segment Occurrences (i.e., non-repetitions and first occurrences of repetitions). Using this shortened workflow, you create a view of First Segment Occurrences on the basis of selected documents, translate this view, and when you have thoroughly reviewed the translation and confirmed the segments to the TM, you pretranslate all documents from the TM and finalize the translation. Sounds easy, right? Well, it's not quite easy if you factor in such things as sharing with colleague translators, review by an external proofreader, and several other things.
If you want to share your view of First Segment Occurrences with a colleague, you need to export the view to a bilingual format (MQXLZ or XLIFF) and send it to the colleague with a note like 'Translate from segment X to segment Y'. You can also optionally lock segments you don't want your colleague to translate before exporting the view, and unlock them afterwards.
When your colleague receives the bilingual file, he/she imports it into their CAT tool and translates it. There are three important limitations here:
1) The translator does not know the context of each segment. Even though all segments appear in the original document order, there are no repetitions, which might be the actual context for a given segment. Without seeing the context, it may be difficult to translate a segment. To make matters worse, if the translator tries to find the segment inside the document to check the context, there may be several occurrences of this segment (because the view contains only first occurrences of repeating segments), so the translator may not be able to provide a context-sensitive translation, or warn the main translator that other repetitions of this segment require a different translation based on context.
2) Even if you send the translator an MQXLZ file which should contain a preview, and the translator also uses memoQ, he/she will not see any preview. This is due to a memoQ limitation related to exported filtered views.
3) If the translator decides to join or split segments, they will not be able to do this because this capability is disabled for filtered views in memoQ.
They may, however, be able to do this if they use another CAT tool.
After the external translator translates their part of the view, they create a view of the segments they translated, export this view to a bilingual file and send this view to the main translator who holds the master file. The main translator imports this view as a new document (memoQ will not update the original view of First Segment Occurrences because the second translator created a view rather than exporting the whole document). Then, the main translator reviews the translation and saves each segment to the TM, and then pre-translates the First Segment Occurrences view from the TM. There are some things to note here:
1) The main translator must use 100% setting for pre-translation because the context around the segments in the document received from the second translator is different compared to the context for segments in the master view of First Segment Occurrences (because the context in a view is defined by the actual context in the document from which the view was created, while the context in the document received from the second translator is defined by segments preceding or following a given segment).
2) If the second translator joined or split segments in another CAT tool, the main translator must perform the same, otherwise there will be no match from the TM. Because joining and splitting in a filtered view is not possible in memoQ, the main translator may need to copy/paste parts of the segments from the document he/she received from the second translator.
On small projects, if you share work with a limited number of translators and the source documents do not exhibit segmentation problems, the points I mentioned above should not present any significant difficulties. As the number of documents and amount of repetitions grows, the project may become unmanageable if the above shortened workflow is used. Our workflow allows to avoid all the above issues because:
1) You export individual documents with locked repetitions, and can then update your memoQ project from bilingual files received from other translators.
2) Because you export individual documents rather than filtered views, the other translators can join/split segments as they like, which is critical if documents were not formatted well or if they were not segmented correctly during the project preparation stage.
3) Because you export individual documents rather than filtered views, the other translators can see document preview if they use memoQ.
4) The other translators can see the context for the segments they translate, even when this context is locked because it represents repetitions.
But does it make sense to use the shortened workflow if you are the only translator on the project and don't need to share work? It does, but you need to consider the following restrictions:
1) As explained above, it is impossible to join/split segments in filtered views, so you need to segment the project thoroughly before you create the view of First Segment Occurrences. If you later need to join / split segments, you can actually remove the view, join/split segments in the original document(s) and then create the view again, but this will take additional time and may need to be done several times.
2) During translation, you can only determine the context of a segment based on preview, since the view of First Segment Occurrences may not contain the full context of a segment due to repetitions. Therefore, make sure that memoQ creates previews during document import.
3) If you work with an external proofreader (some of whom prefer to use bilingual RTF format for proofreading), keep in mind that the proofreader may not be able to see the full context of a segment as they review your translation of First Segment Occurrences, and may find it difficult to find the segment in the original document for the reasons mentioned above.
Again, our workflow is free of the above limitations.
I would like to note that we decided to not include sections on sharing work with other translators and on external proofreading to keep the workflow shorter. However, if there is enough interest, we can include these in a future revision of this workflow.
Best regards,
Stanislav