Re: Idea: each FAQ as its own xml file
Michèle Garoche <[email protected]>
| Newsgroups | gmane.os.apple.fink.i18n |
|---|---|
| Message-ID | <[email protected]> |
Le 18 août 2004, à 18:32, BABA Yoshihiko a écrit : > On 2004/08/18, at 10:07, Michèle Garoche wrote: >>> No one worries English - English discrepancy. >> I worry sometimes, that's why I said it is not the good way to split >> the files. There are already discrepancies between faq, user's guide, >> installation, usage, readme. > That is a problem, but a problem we are not talking about here. Yes, certainly. But it has a huge impact first on the amount of docs, second on translations. Third, it seems discussions of some problems are systematically avoided. > I am sorry I did not explain well. As you say, it will not make is > quicker to translate. My idea is, if translated version does not > exist, the new system will put English version instead. While the > translators are working, the users can see English version. That's what I've done manually. I agree that it would be perfect if it can be done automatically, which I doubt since it implies you know exactly what's the translator's state of works at a given time, and you cannot know it unless you hack his system (hopefully you don't want to do that) and provided it is connected at any time you want to hack it. I explain. Say at time x, faq 1.1 is translated into Japanese, you add faq 1.2 to faq 1.3 in English (say that faq has only 1.1 to 1.3 items to simplify). Meanwhile the translator translates faq 1.2 at it is at time x. Someone changes faq 1.2 in English at time x+1, say now it is faq 1.2-2, with your system you change faq 1.2 in the Japanese version to faq 1.2-2, since the translator has not finished to translate it, so not comitted it and for you it is as if it was not translated at all. At time x+2, the translator comitted faq 1.2 in Japanese, and that is where the fun begins, a big mess in his file and it is not even aware how this has happened, since it is completely hidden in the perl process on the web site. Then good luck to put in it in order. If you tell me here the translator should update his cvs before and after, that's what I've already told. And this is the only reasonable way to ensure that the translation is up to date. Notice that this is not written in the how-to. I say that not on pure hypotheses, but I've sufficiently at numerous times put in order the files, either French or English and at that time, it was more easy because there were traces of what happened. If you ever hidden this, no hope to understand what is wrong, either a process on the web site, a process on the local tree, a text editor which is not adapted to the job, or a person (documentarian or translator) who made an error. >>> Some languages have not been maintained for a few months. Would be >>> a few years. Spanish readers will never know there are more FAQ >>> only if they read English pages. Don't forget that if Spanish readers can read English, they are likely to read the English pages, but if they cannot, what is the utility for them to have a part of the FAQ in English? It is the same as if you told me read this page, it is complete, the first part is in French (ok, that I can) and the second part is in Braille (totally unreadable for me). You're happy because the page is complete, accurate and up to date, but I'm in the same state as before, I can only read/understand (I'll guess this is the purpose of the document, or am I wrong here?) what was before, i.e. the French part. So for me, here happiness is more self-satisfaction than other things. Did you ever ask you why there are voluntaries, but so few remain? I see here these reasons: - no presence on the mailing list of the team project leaders to guide the newbies - no understandable how-to: instead of many pages, one or two are sufficient with accurate 1 to 10 points, and a full list of the files to translate - no quick help to access cvs and to work around some problems in committing etc... >> So, make a call to translators. I've already changed some things in >> Spanish, the Spanish team leader stopped me telling he would make it >> in the weekend (on July 15th). Translators are human beings, they may >> have something more urgent to do or they may be on holidays or ill, >> etc... > It does not seem to me sufficient because some languages have been > left for a few months. To me neither, but I don't want to write down the reasons which come firstly in mind. It could be not very fair to anybody (team leaders, team language leaders, translators, documentarians). But it is not so difficult to find them. And anyway, the endless search on culprits/reasons/responsibilities/etc... is pointless, since it does not make progress things. One has simply to take things are they are, and build systems taking that in account, not dreaming of a perfect world and hope that the perfect system will fit to a real world. >>>> Then, there is another problem splitting the files, the print >>>> version has chance to be only related to the small file, hence you >>>> loose the benefice of searching the whole faq, which is really the >>>> only way to search correctly the documentation. >>>> And if you are about to make this change (as alas it seems), I >>>> don't see why it would be restricted to FAQ, user's guide is also >>>> huge, as well as packaging, etc... Almost any file in xml is huge. >>>> Only the files in web are small. >>>> The dichotomization of things is not always the better way to go. >>>> In this case, it's more like hiding the forest behind the trees. >>> Yes, there still have many things to discuss. I don't think we will >>> change immediately. >> I wonder sometimes why I discuss, because this is not a true >> discussion. You have already decided to make the change, so later or >> sooner it makes no difference. > Because the discussions will make the systems better and better. Provided that things are not already decided, that all arguments are really examined, yes. I'm not sure it is the case here. I'm prone to think the contrary. Michèle <http://micmacfr.homeunix.org>
PGP.sig
(application/pgp-signature, 186 B) - not displayed