Friday, April 28, 2006

Translating contents for small Wikipedias

Well, often it makes a lot of sense to translate articles for small Wikipedias. There are various reasons - one of them for me is: fast contents creation in order to allow other people to go on working on what I put there. Now, when it comes to whole articles people still often ulse the slowest method: open a Wikipedia page and copy the contents pasting it to another Wikipedia page where they then translate it step by step. A faster approach, using a good glossary, is the work with a CAT-Tool like OmegaT, but you still have to open the wiki page, copy the text, paste it in the text editor - or in OpenOffice.org, save it, translate it with the CAT-Tool and then open the .odp file and copy and past the text to its destination page. Now this copy and paste takes some time and if you do it frequently you will note how fast you reach an hour that you loose in only copying and pasting ... time ago there was that proposal of a wiki-read - translate - wiki write feature for OmegaT. It would be great to have it - simply to save so much time and to be able to do much more for the wikipedia instead of handling copy/paste mechanisms.

Then there is that part I mainly do on the Neapolitan wikipedia I mainly translate calendar pages there - or better actually only the "born" people. I did this for approx. 15 pages up to now and today I noted that 80% of the work is already done, because most people listed are those we all know: writers, composers, politicians etc. OmegaT is particularly valuable with this thanks to its extraordinary flexible segmentation rules - they allow me to strip out only the part to be translated since the name of the person remains - only the profession or what that person was known for needs translation. The segmentation rules are based on regular expressions and compared to my first calendar page they make me save 80% of the translation time :-)

This is simply something we need.

A further advantage: it could happen that on certain wikipedias people simply decide to update certain articles always according to their original source ... for certain stuff that would make sense (example: science) - this means that if an article is updated and you want to update the translated one you just have it run against the old Translation Memory you built and all that remained the same will show up and does not need to be translated anymore.

A further advantage is terminology consistency and style consistency when several people work on similar or related articles. They can exchange translation memory and search for terminology in there (besides exchanging their glossaries that one day anyway will come from WiktionaryZ). This means better quality and faster translation because over time more and more terminology will be in the Translation Memories and glossaries and so people don't need to search in further ressources.

All in all having that wiki-read - translate - wiki-write feature would be a huge advantage. I would simply love it.

No comments:

Khalil Gibran über die Musik

Die Musik wirkt wie die Sonne, die alle Blumen des Feldes mit ihrem Strahlen zum Leben erweckt. ( Khalil Gibran ) Image by Pete Linforth fr...