November 12, 2008

Making life easier for new localizers

When you look at recent posts in the Mozilla l10n space, sooner or later you'll find out that a lot of localizations only exist for Firefox, but not for the comm-central application Thunderbird, Sunbird or SeaMonkey. If you look at the numbers, you'll see that Firefox 3.1 currently has 65 localizations compared to 48 localizations for Thunderbird 3.0, 36 localizations for Sunbird 1.0 and 18 localizations for SeaMonkey 2.0.

This is not totally surprising, since Firefox has many more users and much more buzz than all three other applications combined. And even for locales that support more than just Firefox, Firefox is the natural leader in terms of localizer attention and status, because of its market share and popularity. This is nothing to complain about since people generally are more likely to work on something that makes a big difference and localizing Firefox into e.g. Vietnamese makes a much bigger difference than localizing Sunbird into Vietnamese instead.

For us in the Thunderbird/Sunbird/SeaMonkey communities this means that we need to adjust to this fact. To be more precise we need to make the actual localization for people as easy as possible, because when push comes to shove our apps will always come in second, third or fourth.

Most of us obviously can't help a new localizer with the basic translation parts, since in most cases we don't speak the language, but what can be done is to make all the other aspects of localizing a Mozilla application as easy and as less burdensome as possible. This means
  1. reducing the bureaucracy around the l10n processes
  2. providing better tools for localizers
  3. communicating clearly and effectively with localizers
  4. making sure, that localizers can concentrate on what they are good at, the actual translation/localization part
On the first item, I think we are doing pretty well, since we do not require nearly as many approvals, reviews, etc. from localizers as Firefox does. I hope we can continue to keep our processes lean as we grow our communities and get more more localizations.

The second item is very well covered by the growing l10n team at Mozilla Corporation. Especially Pike and Gandalf are currently doing great stuff in the tools space, that will make life easier for many localizers.

The third item is hard for me to comment on, since I'm doing the l10n communication for Thunderbird and Sunbird and it is hard to judge yourself. I haven't heard any complaints yet, but maybe I just missed it. I certainly hope that people will give me constructive feedback if I screw up.

On the last item, we made a few huge steps lately. For two weeks various people (including me) have done a script-based analysis of our codebase in search of unused strings. Unused strings are bad for new localizers, because you need time to translate them, which could be put to better use. We recently fixed three bugs or are in process of fixing them, which removed (or will remove) a total of 670 strings (219 for Thunderbird, 127 for Thunderbird and SeaMonkey, 324 for SeaMonkey. This is a lot of stuff (for Thunderbird roughly 1/8 of all localizable strings) that a new localizer will no longer have to wade through.

No comments: