On Mon, Apr 21, 2003 at 01:57:32PM +0200, Kaare Olsen wrote: > On Mon, 21 Apr 2003 10:56:25 +0200 > Frank Lichtenheld <[EMAIL PROTECTED]> wrote: > > > > Is the German translation based directly on the Swedish original? > > > > No. But if I base the translation check on the english one, and the > > english one is out-of-date, the user of the german page gets no > > warning (neither get I). > > And now you can't trace updates to the English translation, the one your > translation is still based on. What if there are errors (i.e. an > incorrect link or a missing paragraph) in the English translation which > gets fixed? You wouldn't know since the original could be correct and > hence not updated.
Ok, I see your point here. So it's a matter of priorities: Is it more important that I am informed on changes to the English file by the stattrans.pl statistics or is it more important that the user sees that he gets a page that is outdated. This not so important for this particular page but I also changed the translation check header in german/international/Russian.wml where the English translation is very old. The English user can recognise this because of the translation check, but the German user got no hint since the German page is up-to-date with the english one. This was not acceptable for me. I have other possibilities to recognise changes to the english file (like debian-www-cvs), too. > [...] > > Which one I use for translation, is an other matter :) > > Then at least add a comment to your translation, saying that it is > actually based on English version x.y, otherwise you're the only one who > knows. :-) But I still think that you should point to the language your > translation is based on. This is a good point. But it is clearly only a workaround. Perhaps I will really try to implement the chain check I spoke about. Sounds like some nice Perl hacking :) Frank -- *** Frank Lichtenheld <[EMAIL PROTECTED]> *** *** http://www.djpig.de/ *** see also: - http://www.usta.de/ - http://fachschaft.physik.uni-karlsruhe.de/