On Wed, Jul 07, 2004 at 09:15:37PM +0100, Peter Karlsson wrote: > > That kind of intelligence does not have to be artificial. What about the > > original writer estimates whether his modifications are merely cosmetic or > > require a new translation ? > > That might be nice, but I don't think such a scheme would actually work. It > would require the original authors (which are many) to agree on how to > handle this. There should be clear policies to avoid this: -Developers' Reference</a>. Adam Di Carlo will work on a +Developer's Reference</a>. Adam Di Carlo will work on a forcing a new revision of a translation. Somebody here made a small change and forced 8 different translations to revise. That's 8*5 = 40 minutes (under Gerfried's estimate, which I don't agree with) wasted of other people's time for a simple change. It is simply not acceptable. I can find many other examples of similar wrong behaviour. > Yes. That's why there is a six month's grace period. If you haven't got time > to check your files for *minor* updates in six months, you aren't doing your > job, volunteer or not. I'm sorry but you don't get it. You don't really understand, do you? You think that translation removal is a way to punish the translation team for not doing it's job? You are actually punishing the casual web reader that stumbles into a foreign (and sometimes meaningless to him) web page when a translation that was "good enough" could have served him just fine. > > So you are actually saying that a translator can actually bypass the > > modifications and resurrect a file unmodified to satisfy the needs of an > > obsolete verification model ? > > The translator can resurrect the file, bump the version number and commit it > again. If he just resurrects it without updating the version number it will > be autopurged again next week. It is not that easy to detect which files were removed unless you dig in the Attic in the CVS server directly. Or you happen to have a CVS checkout before the massive removal was made. If you can tell me another (simple) way my ears are listening. > If the translation teams can't handle the load they should either get more > people or focus on a smaller number of pages. I would like you to explain me how can typo changes in the News/weekly hierarchy be reason enough to remove perfectly correct translations. I take News/weekly as an example of when _not_ to remove translations. Any changes there can only be typo fixes (or cosmetic changes as above) which should not force a removal, regardless of wether the translation team checked on them or not. > > We are not talking "ignorance" here but "irrelevance". And "irrelevance" > > is definitely an excuse. > > If the translations are irrelevant, then don't do them. Concentrate on the > important pages. There's no need to translate all of the pages. Just because > I haven't got a life and thus have the highest number of translation for any > single language for the Swedish translation doesn't mean that all the pages > are important to translate (I know several which aren't, and which I'm > thinking of deleting because *I* can't keep up with them). The funny thing is that your removal also affected content that should _not_ be dynamic. We are not talking about content that was substantially modified, we are talking about content that was written (and translated) years ago and now has been undergone cosmetic (in the english speaker's opinion) changes. The translated content was still valuable, and you pushed it aside. > But as I said earlier. If anyone can come up with a better system that is > guaranteed to work, please do. Keeping dead translations around is *not* a > better system. I don't see why not. We shouldn't worry about a given file that has not been updated for a while. You can put bigger warnings if you like but you don't speak for the Spanish-speaking population of the world. That language speaking population is a rather big one, and just like the Chinese-speaking (which is even bigger) population is not guaranteed to be able to make use of an english page. I can give you a zillion examples of people that would rather have an outdated translation than no translation at all. In your "holier than thou" attitude you are only punishing our users to no benefit at all. As long as the translation-check mechanism is used when typographical or cosmetic changes are done to pages (and for the look of the spanish webpages translation I would say that it's one of the primary causes of changes, specially under some hierarchies) then your metric of "six months" is bound to fail regardless of the arguments you want to make sustaining it. I think that removing stuff is only guaranteed to be of use when a language, as a whole, is not able to keep up with translations. We've removed languages in the past for this reason, and we should do it in the future too. If you want to do "fine grain" analysis, either you do it byhand (not through a script) or you use a metric that is agreed by all. I know I didn't dispute it a while back (in january IIRC) but I do it now. Actually, a metric I would incline for is (Original version - Translated version) > 20 (or an arbitrary, but high number). That guarantees that there have been much more changes than cosmetic ones. Best regards Javier
Attachment:
signature.asc
Description: Digital signature