-
How to build a language pack…en Español
Ever wonder how to create a language pack? Here is a video from Ricardo Meza from our Mexican community describing how to do it….in just over 7 minutes.
The video is in Spanish, so if you don’t speak it, you’ll have to get some help. Ricardo sent it to me last night after we chatted for sometime about our localization community in Mexico. If we were to build and release es-MX, 97% of the country’s Internet users would be able to surf with a browser in Mexican Spanish along with web services relevant to the Mexican browsing/user experience.
Ricardo and many others in Mexico are excited to make this happen. Interested in helping? Comment on this blog post.
-
Mozilla’s L10n Dashboard
Check out this tool from Pike. He calls it the l10n dashboard. It’s a great snapshot into the state of Mozilla l10n community. I’d say it’s a work in progress with a way to go, but so far, it’s very well put together.
As our team tries to improve efficiency and outreach/service to localizers in our l10n process, this tool will greatly help in how we focus on and respond to community needs related to localization.
My best suggestion is to click through and start playing with it, asking questions along the way on this blog. We can help explain what might be unclear.
But, here’s a brief set of some of the things you can do to get you started.
Main Page
The landing page is all 65 localizations that are presently a part of Mozilla’s build process. This does not mean we ship 65 localizations…it means that they are in CVS and can be built nightly and tested. Nota bene, this does not include language packs.
Top navigation of the table
Each column header is sortable. (Do we need to make that more clear? Not sure if it’s clear to me at first glance that you can play with this chart.) The first column is the language code and it is sorted alphabetically, with af (Afrikaans) first.
The second column is titled “Tree” and two values: “trunk” and “incubator”. Trunk means the builds are on the Firefox 3.0 trunk in CVS. Incubator means that we need to migrate the locales over to 3.0. When I saw this, I immediately began to ask, “What have we done to reach out directly to those locales in the incubator to see what we can to do to get them onto the trunk?” We’re working on it…
The next column is a “percentage complete/changed”. You might ask, how could anything ship with anything less that 100% complete? Or, what does “complete/changed” actually mean? Let me give an example. Someone in the es-AR localization may not have a translation for a country name of an island located in the Pacific ocean. That string will not be changed. But, it doesn’t mean that es-AR cannot ship a near-perfect version of Firefox for its users. Does that make sense? There are lots of entities in our code that just are not going to have a translation in some locales.
The “missing” columns shows you all the locales that will not build because they have missing pieces that need translation. If you sort on missing to show all the builds that have missing values, you’ll see that it is the same number as the “failure” link, which is just under “success”. This means that those localizations teams are working very hard to get their builds ready to ship in 3.0.2 (or later)…or that they have yet to be migrated over to Firefox 3 and still haven’t finished localizing Firefox 2.
Other stuff
On the right side, you’ll see a bunch of boxes. Click around with those and you’ll start to see how we are trying to manage our own capacity and work flow. Sadly, we are not automatons (Ken Kovash is the only robot at Mozilla), so we need to figure out where we can help our localizers.
Click “reset all filters” at the top. Then, in the bottom box on the right hand side called “Shipped in…”, click on “Missing this field…” You’ll see 17 localizations that have not shipped in Firefox 3. Bingo! This is the focal point. How do we get these 17 into Firefox 3?? We’ve reached out directly to a lot of these in the past four weeks, making time for personal IRC meetings, emails, IMs, etc. Check out bug 442935. After these meetings, we landed 8 new languages with this bug to the build process. In total, ten new localizations have been added in the recent past. We’re making progress and have a lot to do like working with communities to get the localizations ready to ship. If you’re in the dashboard’s “list of 17″ and haven’t heard from us, we’ll probably ping you soon. But, don’t hesitate to contact us because we want to hear from you.
Pike, Mic, Pascalc, Chofmann, and I talk almost daily on how we can improve this dashboard. It’s really a great tool and much, muCH, MUCH thanks to Pike and Pascalc for putting this together. If you have ideas, I’d really encourage you to participate. This is a perfect chance to get involved and make your ideas a part of the future of Mozilla l10n.
-
Frequently Used Entries for Localization (FUEL)
Rajesh Ranjan is one of our lead localizers in India. Aside from working to complete the Hindi translation of Firefox, Rajesh does a lot of community organizing and outreach to open source enthusiasts around localization efforts.
He recently held a workshop to disuss a new effort called FUEL: Frequently Used Entries for Localization. Rajesh describes FUEL in a blog post with the following:
“…all major desktop related entries [needing localization] appearing on menus and sub-menus are not more than five-six hundred. So if we move to standardize a mere 500-600 entries and the process is backed by the active localizers and entities who get benefit from localization then we can make a successful move against the problem of standardization and inconsistency in software translation. This is the main idea behind FUEL.”
To learn more about FUEL, please read this post by Rajesh. And, here is one more link about FUEL.
-
Firefox 3 Sinhala Release on Swarnavahini News
Danishka is our lead Sinhala localizer in Sri Lanka. In addition to creating a localized version of Firefox 3, he’s been doing some great evangelizing about Mozilla in his home country. Check out this video on Swarnavahini News. (If you don’t understand Sinhala, this video is gonna be rough. But, you’ll still see some noticeable logos and hear some familiar names.)
Firefox 3 Sinhala Release on Swarnavahini News
Danishka, how’d you do it? It’s so impressive that you localized the browser AND have been creating such great buzz in your country. Very cool!!
More news on impact out of Colombo.
Many thanks to our friends at the Information and Communication Technology Agency of Sri Lanka (ICTA) and the University of Moratuwa who helped Danishka in this great effort.
I would love to hear any thoughts from our community in Sri Lanka. How you are expanding? How did you find the localization process? What tools did you use? Please give us your comments. I know others would be interested in reading.
-
Reviewing l10n goals
Last week, a subset of the l10n-drivers team met in Paris (my pics are coming) to discuss many things related to l10n, including a review of goals that had been set last November before the lead-up to the release of Firefox 3. The goals were ambitious and listed many ways in which the team could create positive impact. Part of our work week was to review these goals and find out what we did well, didn’t do well, and could do better as we create goals for the upcoming quarter.
One thing to note for this post, I’ve tried to distill lots and lots of goals (sometimes, they were simply tasks) into larger buckets that represent common themes among the ideas that came up. I’ve tried to list what was done well and what needs improvement for the goals that were created.
Goal 1: Create a finite schedule that localizers could follow as they worked up to the release of Firefox 3 and Thunderbird
How did we do? Well, we did the following:
- Improved the build infrastructure using L10n server
- Created a dashboard that shows state of locales and outstanding bug
- Started filing tracking bugs for each locale interested in shipping FF3
We could improve by getting better at time management of certain aspects of localization that become obsolete (specifically web stuff), so our localizer community does not working fruitlessly to translate something that is then not used.
Goal 2: L10n team implements more strategic initiatives that benefit the Mozilla Community
How’d it go?
We worked hard on the following:
- Proactively developing and supporting community
- Contributing to marketing/analysis/metrics teams
- Create tools for the l10n community
- Improve communication about the localization process so people can plan more efficiently
Overall, we did pretty well here and can continue to improve. For community development, we had hoped to find new, motivated individuals to join teams and help lead, create opportunities for localizers to grow into new marketing and PR roles if interested, and to encourage localizers to help Mozilla understand local initiatives, goals, and even trends taking place. Essentially, we want to build good relationships with localizers. The trouble with this is that it’s hard to measure and it’s an ongoing process. So, we couldn’t really find concrete examples, but are starting new initiatives like the Community Pack, which I’ll describe below.
Regarding metrics, Ken Kovash’s blog has a lot of information about locale specific use of Firefox. We hope to grow more here and do more with Ken so our localizers have more information about the impact they are having.
Tools development also had some success. We are made a grant to Translate.org.za for their work on Pootle. Wil Clouser and Dan Schafer have been working very hard on Verbatim. We also held a tools BoF at FOSDEM. We chatted with Narro Project several times and Narro opened its code for all to see and contribute. More work will continue on tools development throughout the coming year.
From a QA/testing perspective, we did the following:
- Provided better, more timely feedback when comparing the en-US trunk vs. localizer’s code
- Used Axel’s compare-locales tool to test more bugs
- Developed a screenshoot tool which goes through preference dialogues and enables a bunch of screenshots and tell everyone what the dialogue sizes are. (By the way, Pike ran all this manually. What a job!)
We can improve by providing more support to new localizers. Zbigniew Braniecki is spearheading one idea that he calls the Community Pack that will be a list of resources available for the community. We also need to enhance our testing tools for localizers.
Goal 3: Mozilla’s localization-drivers team define its role and scope in the Mozilla Corporation and in the Mozilla Community, answering the questions “What are the l10n-drivers? What do they do? And, how can they help us?”
How’d we do?
We probably could have done a lot better here, but the good thing is that the team is focusing its goals for the upcoming quarter and, through my blog and other outlets, we’ll be much more transparent and communicative about what we are doing.
I mentioned it last post…you may have noticed that I’m writing more and more about l10n. Well, there is a reason. I’ve been asked to change my role at Mozilla and take on more leadership with our localization efforts. I will continue to work with community to provide very leveraged support with the Community Giving and Empowerment Program. However, that experience and my involvement with the Evangelism team gave me a lot more exposure to things like the l10n community. Now, I’ll be working on the l10n team to improve our process, blog more to provide transparency and showcase ideas and progress, and work hard to build new l10n communities while servicing our existing localizers who do so much. Hooray! I couldn’t be more excited. And, I am so honored to work with all of the localizers in our community who make it possible for our software to ship in so many languages.
community website
