Monthly Archives: May 2008

Two paradigms: Gettext and Mediawiki

Gettext starts from quite a different perspective to i18n. Especially it differs in who should do the ugly work. The thing is, Gettext tries to hide the i18n from developers, while the system we build in Betawiki aims for minimising the work translators have to do. These two aims produce systems that are different and it needs some thought how to combine them together. Fortunately these two aims aren’t entirely incompatible. I have to say that hiding i18n from the developers has its good and bad sides, but I’m not to judge whether it has more good or more bad.

Paradigms aside the main difference seems to boil down to tracking changes to messages. Betawiki does it, and it is easy because every message is identified by a unique name. Gettext doesn’t really, it just prefills the translations for new and changed messages by guesswork.

We use MediaWiki pages, which have a concept of unique name. So obviously I need to generate some kind of unique names for the messages in Gettext files. Maybe hash of the contents and context, which is the Gettext definition of uniqueness. Not pretty, but as the developers aren’t forced to name the messages, there is probably no way to get meaningful names.

That should be a start at least. I have been fiddling with the code trying to separate file format support to own classes from other code, but I’m not yet happy with it.

I also hope I figure out some clever trick to track messages changes from .po files to keep more history in the wiki. It may be that i could use Gettext guesswork algorithm to some extend, but it may also be that it is not worth it, due the nature of Gettext. In any case we have the history for changes in the translations.

Memory optimisations

Yesterday (or in the midnight hours) I finally committed a patch to MediaWiki’s message cache. Betawiki uses MediaWiki in a way that puts a heavy pressure on the message cache. While normal MediaWiki installations have maybe dozens or few hundreds of customisations to MediaWiki interface messages (pages in MediaWiki namespace), Betawiki has hundreds of thousands of messages in hundreds of languages

The amount of messages that needs to be cached effectively is really in a different decade. Normally those messages take maybe few hundreds of kilobytes in PHP’s serialised format, stored in the database or in memory cache. In Betawiki all messages together would take about 23 megabytes! It is clear that loading and handling such a big blob is not going to work, especially when it is needed on every page request and needs to be updated on every change to the messages.

Some time ago we started to hit the memory limit we have set for PHP requests. I made some hacks to the code reduce the burden—but those were only hacks. Before this patch we basically stored only customisations to be used for Betawiki itself and skipping message cache updates totally, so it would only be updated after a timeout.

This was far from an ideal solution. The message cache was caching all the other messages individually. This is of course waste of memory and more importantly fragmentation increased a lot and request per second to memory cache (we use APC in Betawiki) sky-rocketed to thousands per second.

What made me hesitant to commit this patch was, that I needed to update code paths we don’t use in Betawiki, and thus wouldn’t get a much real testing. At the time of writing this message, it seems to be live on the servers of Wikimedia Foundation and is not reverted or got any comments so far, so it probably isn’t totally broken or unacceptable :).

What the new patch actually does, is that it adds a new configuration option, which when set to true will split the cache to smaller caches that contain messages for one language only. This greatly reduces to memory consumption, as only a couple of languages needs to be loaded in normal use. Full localisation of MediaWiki and all supported extensions takes from 500 to 800 kilobytes, depending on the script. The default setting for the new configuration option is false, which should result behaviour identical to the old version. I also added more comments and standardised the names of per language memory cache keys.

This will not solve all memory use problem in Betawiki, but is big step to keep it running efficiently, and with as few hacks as possible. Custom hacks are bad because they add maintenance burden and prevents others from creating a similar setup easily.

Of course the amount of messages will only grow in the future. To tackle this I have planned to move non-MediaWiki related messages to a another namespace, so at message cache will not handle them at all.

Betawiki status report

It is raining again—or at least it would be nice if it did. Betawiki has had some nice new progression in the spring. Aside from the general growth in translators, page requests, translation and languages, the community itself has evolved.

They have created a news letter that is sent out once a month at most. Some translators have started to suggest enhancements to the messages, for example if it is missing plural handling or bad wording that is hard to translate. Also some of our projects pages are being translated, even though the process to do so is a bit awkward.

As a platform we have adopted one new external project Word2MediaWiki plus, which converts word documents to wikitext. New extension named Babel—used by the users indicating what languages they speak and how fluently—is in development, and Betawiki has helped by providing translations and by acting as a test platform. Babel extension is developed by MinuteElectron. Let’s hope it will soon get ready for use in Wikimedia projects.

Also the first external project in Betawiki—FreeCol—got some revitalisation. I have agreed with Michael Burschik that I commit language updates from Betawiki once or twice a week. As always, faster integration cycle helps in testing the translations and messages themselves before release. Well, not everything is great and perfect yet. FreeCol development is active in the trunk branch mostly, while the translation in Betawiki are for 0.7.x branch. We support branched translations for MediaWiki and it should be possible to do so for FreeCol also. There are some jumps and hops to go trough, so it hasn’t been done yet, but should be quite easy. Also, we currently can’t generate statistics for FreeCol, but that will be fixed too.

I’m quite happy how the different work tasks are spreading out. It’s not only one man’s project anymore, and normal things go forward even if I’m not there every day. It leaves me more time to actually make it better than just run the whole project. :)

-- .