Tip for huge word-files
Thread poster: Ivan Eikås Skjøstad
We have all worked with huge word-files with a lot of formatting with the result that Word crashes....
I am working with a 200+ pages document now, and when using Wordfast or Trados the document was slow and crashed due to the size.
What I figured out was that the CAT-tools seem to be fragile during the segmentation process, and I did a "Translate" in Trados 6.5 and ticked "Sement unknown segments". Now my document is 400+ pages, but very stable.
To make sure everything works even faster I have turned off term rec, but if the document was a bit smaller I think that it should work well with term rec turned on as well. Update: I have turned it on again, and it works well even for this large document.
One could say that it would be easier to split this document in two or into four smaller parts, but the formatting was too complex.
Just a tip!
[Edited at 2005-03-09 11:37]
| | Jabberwock
Local time: 07:01
English to Polish
| Export works too... || Mar 9, 2005 |
What I often do is to analyze the whole package and export unknown segments into an RTF file. Depending on conditions sometimes I pretranslate it (with segmenting), sometimes I don't.
Then I translate that file and use the resulting TM for final translation/revision of the original document - sometimes I use automatic translation, most often I open each segment again and read the proposed segment.
- It is pure text document, easy to handle.
- All segments are in one document. Useful, when translating many similar documents.
- If I pretranslated with segmenting, I can change repeating portions of the segments without actually opening them (replacing only unhidden text).
- All changes done to translated segments may be (and have to be!) included in the database with the clean up command.
- If a subcontractor translates the RTF, it is very easy to align it with the original (segmenting only on hard return).
- It is easy to forget that it is _not_ the document that is translated, only the TM segments. Therefore changes made only to the document (multiple translations of a single segment, for example) are lost!
- With repetitions the source text has to be compared with the original (printout). The repeated segments are simply not there, which may disturb the context.
- Tags require special attention. It is easy to lose them, especially for segments where only tags differ.
- Joining and splitting segments is out of question, as they won't be recognized in the original document.
Considering the above, I don't recommend trying it out on a large project, or even any project that may cost you money. It may work for some documents, it may not for others. You have to know exactly what you're doing.
| || || |
Thank you for your comment. This is interesting reading. I will try this on a smaller project some time.
A good tip, but still I would translate huge doc files like that with Déjà Vu X or SDLX. They import/export the text and I have never seen them crash.
And if you can save the document in rtf format, I bet it would be more stable to translate it in TagEditor than in Word.
| | Jörgen Slet
Local time: 08:01
English to Estonian
| I wish CAT tools were compatible with better editors || Mar 10, 2005 |
I wish CAT tools were compatible with better editors with more scalable architecture, but I know that will never be. I used to use Framemaker to handle all my documents a few years ago and it could handle very massive files and still remain very snappy and stable.
To report site rules violations or get help, contact a site moderator:
You can also contact site staff by submitting a support request »