So.. whats the status on this? Noone cares? Are we disabling most of the translations (to begin with) and take it from there?
-Hannes On Nov 11, 2007 3:13 PM, Hannes Magnusson <[EMAIL PROTECTED]> wrote: > On Nov 10, 2007 5:20 AM, Philip Olson <[EMAIL PROTECTED]> wrote: > > Hello everyone, > > > > As most of us know, we have many outdated translations... so let's > > discuss it: > > > > A) Critically old files: > > > > Many translations contain critically old files that should be either > > updated or offline. Some ideas that deal with these are: > > > > - Have the build system either not build/show them, or insert huge > > warnings (for users) > > - Add revcheck[1] tools that list all critically outdated files (for > > translators) > > - Better define what it means to be critically old (for all) > > Since the "revision element discussion" didn't get us anywhere we have > no real why of figuring out why files or outdated or determine if a > file is really critically outdated or not :( > > > > B) File ownership: > > > > Translators typically insert maintainer information within each file. > > If a translator suddenly becomes inactive, these files essentially > > become unmaintained yet remain owned. I don't know the extent of this > > problem but can only assume it causes delays. Should we worry about > > allowing active translators to update any file... especially for > > critically old files? They (some) do anyways but let's make something > > official. > > This is somewhat related to general crediting documentation writers > and the changelog discussion we've been talking about (offline). > I think however the main reason for the "file ownership" is so the > translators easily check the files they are interested in translating > and update them if needed without needing to scan the entire tree for > changes. > > > > D) Outdated translations: > > > > We have several translations that haven't [really] been updated for > > years, and it goes without saying that this is bad for everyone so > > let's make a plan. Here's one, let's discuss it: > > > > 1. Designate the deadest of the dead as INACTIVE_LANGUAGES via > > phpweb. ~18 of them. This means they won't show up via the select box > > at php.net, nor be selectable via my.php. > > (and the user will not be automagically redirected to the translation, > even if they send out ACCEPT_LANGUAGE header for that language, and > the translation is not listed on php.net/docs nor > php.net/download-docs) > > > 2. Write each list (doc-{lang}) asking if anyone out there is > > listening. If so, discuss the translation. > > So if someone is "listening".. then what? Keep the 3year old language > available? > I'm willing to bet that its way easier to start from scratch for 99% > of these translations. > > > > 3. Alter the php.net 404 handler to work with missing languages, so > > ar/manual/foo.php --> en/manual/foo.php > > That shouldn't be a problem, for _missing_ translations, but for > translations that are in fact online and someone writes a > php.net/full/path/to/file.php.. > > > 4. Implement PhD to build active languages for mirrors rsync. Based > > on INACTIVE_LANGUAGES from phpweb/includes/languages.inc. > > 5. Implement PhD to build all languages, active and inactive, for > > docs.php.net. > > 6. Remove all dead/old/non-phd manuals. For example, kr/manual is > > from 2004. Currently some translations (even en/ within them) are not > > being updated/built. > > Most of the translations online haven't even been rebuilt using xslt, > which I find terribly annoying - especially since there is a bunch of > legacy crap that is in my way and I'd like to remove - and their > layout is totally different from xslt, then adding phd builds on top > of that... its impossible to maintain three "almost like > markup/classes" and expect them all to look alike. > _We need those 18 translations disabled and phd builds pushed out_ > > I have no special feelings regarding removing them from phpweb or not. > > -Hannes >