I am working on a Word document which includes several tables, and I've noticed that segmentation seems to be horizontal (rows) and not vertical (columns).
Because some information is split into multiple vertical cells, the result is that the content of these cells is not adjacent in CafeTran, and can therefore not be joined.
Unless I am missing something, I suggest that it could be more practical to segment these tables vertically and not horizontally within Word (and ODT) documents.
The Word tables generally not spanning into multiple pages, the horizontal context should not be such an issue.
This would also be consistent with the way Excel files are treated (per column).
What do you think?
Conceptually, it is a good idea. However, CafeTran's current filters parse the text as it appears in the parsed file. In certain situations (e.g. with footnotes) the text can be shown in other coordinates of the final view. The above issue with table columns and rows is related. The solution would be to write the parsers from scratch to take into account the actual documents layout in every supported filter, which in practice mean rewrite Word, LibreOffice writer, or to use some third-party libraries. I know that users would ideally like to see the text in CAT tools as sort of embedded in original applications. CafeTran offers a method to translate straight from the original applications via the Clipboard workflow. That is really a viable solution in cases where the real-time layout is essential, and it might work for you with the rows/columns segmentation issue.
Your explanation is much appreciated.
This is a limitation to be aware of when accessing the best way to tackle word processing formats which include rich table content.
Although not ideal, it is still workable to translate some embedded tables this way. Fortunately, human mind is still able to “virtually join” segments. :-)
One other solution would definitely be to use the Clipboard workflow, perhaps along with a way to hide the tables during the normal translation. This workflow looks nice on paper, but I haven’t found a way to integrate it yet in my own way of working. Will try again in the future.
Another idea would be to use the round trip method if another tool offers a different implementation. I will check this, and maybe add a note on tables in the reference file on file formats.
I’m sure there are other ways as well.
Jean: I’m sure there are other ways as well.
I'm not exactly a Word expert (nor a fan), and I never encountered the problem, so I can't test my Simpleton solution, but assuming the "splitting" in the Word table is "soft", can't you get those split fragments in one line, for example by making the fonts smaller, widening the margins, using panorama view?