Pseudolocalization in Firefox

One of the core projects we did over 2017 was a major overhaul of the Localization and Internationalization layers in Gecko, and all throughout the first half of 2018 we were introducing Fluent into Firefox.

All of that work was “behind the scenes” and laid the foundation to enable us to bring higher level improvements in the future.

Today, I’m happy to announce that the first of those high-level features has just landed in Firefox Nightly!


Pseudolocalization is a technology allowing for testing the localizability of software UI. It allows developers to check how the UI they are working on will look like when translated, without having to wait for translations to become available.

It shortens the Test-Driven Development cycle and lowers the burden of creating localizable UI.

Here’s a demo of how it works:

How to turn it on?

At the moment, we don’t have any UI for this feature. You need to create a new preference called intl.l10n.pseudo and set its value to accented for a left-to-right, ~30% longer strategy, or bidi for a right-to-left strategy. (more documentation).

If you test the bidi strategy you also will likely want to switch another preference – intl.uidirection – to 1. This is because right now the directionality of text and layout are not connected. We will improve that in the future.

We’ll be looking into ways to expose this functionality in the UI, and if you have any ideas or suggestions for what you’d like to see, let’s talk!

Nitty-gritty details

Although the feature may seem simple to add, and the actual patch that adds it was less than 100 lines long, it took many years of prototyping and years of development to build the foundation layers to allow for it.

Many of the design principles of Project Fluent combined with the vision shaped by the L10n Drivers Team at Mozilla allowed for dynamic runtime locale switching and declarative UI localization bindings.

Thanks to all of that work, we don’t have to require special builds or increase the bundle size for this feature to work. It comes practically for free and we can extend and fine tune pseudolocalization strategies on fly.


If that feature looks cool, in the esoteric way localization and internationalization can, please, make sure to high-five the people who put a lot of work to get this done: Staś Małolepszy, Axel Hecht, Francesco Lodolo, Jeff Beatty and Dave Townsend.

More features are coming! Stay tuned.

Multilingual Gecko Status Update 2018.1

As promised in my previous post, I’d like to do a better job at delivering status updates on Internationalization and Localization technologies at Gecko at shorter intervals than once per year.

In the previous post we covered recent history up to Firefox 58 which got released in January 2018. Since then we finished and shipped Firefox 59 and also finished all major work on Firefox 60, so this post will cover the two.

Firefox 59 (March)

Firefox 58 shipped with a single string localized using Fluent. In 59 we made the next step and migrated 5 strings from an old localization system to use Fluent. This allowed us to test all of our migration code to ensure that as we port Firefox to the new API we preserve all the hard work of hundreds of localizers.

In 59 we also made several improvements to performance of how Fluent operates, all of that while waiting for the stylo-chrome to land in Firefox 60.

In LocaleService, we made another important switch. We replaced old general.useragent.locale and intl.locale.matchOS prefs with a single new pref intl.locale.requested.

This change accomplished several goals:

  • The name represents the function better. Previously it was pretty confusing for people as to why Gecko doesn’t react immediately when they set the pref. Now it is more clear that this is just a requested locale, and there’s some language negotiation that, depending on available locales, will switch to it or not.
  • The new pref is optional. Since by default it matches the defaultLocale, we can now skip it and just treat its non-presence as the default mode in which we follow the default locale. That allowed us to remove some code.
  • The new pref allows us to store locale lists. The new pref is meant to store a comma separated list of requested locales like "fr-CA, fr-FR, en-US", in line with our model of handling locale lists, rather than single locales.
  • If the pref is defined and the value is empty, we’ll look into OS for the locale to use, making it a replacement for the matchOS pref.

This is important particularly because it took us very long time to unify all uses of the pref, remove it from all around the code and finally be able to switch to the new one which should serve us much better.

Next come a usual set of updates, including update to ICU 60 by Andre, and cleanups by Makoto Kato – we’re riding the wave of removing old APIs and unifying code around ICU and encoding_rs.

Lastly, as we start looking more at aligning our language resources with CLDR, Francesco started sorting out our plural rules differences and language and region names. This is the first step on the path to upstream our work to CLDR and reuse it in place of our custom data.

Notable changes [my work] [intl]:

Firefox 60 (May)

Although Firefox 60 has not yet been released as of today, the major work cycle on it has finished, and it is currently in the beta channel for stabilization.

In it, we’ve completed another milestone for Fluent migrating not just a couple, but over 100 strings in Firefox Preferences from the old API. This marks the first release where Fluent is actually used to localize a visible portion of Firefox UI!

As part of that work, we pushed our first significant update of Fluent in Gecko, and landed a special chrome-only API to get Fluent’s performance on par with the old system.

With an increase in the use of Fluent, we also covered it with Mozilla eslint rules, improved missing strings reporting, and wrote an Introduction to Fluent for Firefox Developers.

On the Locale Management side, we separated out mozilla::intl::Locale class and further aligned it with BCP47.

But the big change here is the switch of the source of available locales from the old ChromeRegistry to L10nRegistry.

This is the last symbolic hand-over from the old model to the new, meaning that from that moment the locales registered in L10nRegistry will be used to negotiate language selection for Firefox, and ChromeRegistry becomes a regular customer rather than a provider of the language selection.

We’re very close to finalize the LocaleService model after over a year of refactoring Gecko!

Regular healthy number of cleanups happened as well. Henri switched more code to use encoding_rs, and updated encoding_rs to 0.7.2, Jeff Walden performed a major cleanup of our SpiderMonkey Intl source code, Andre added caches for many Intl APIs to speed them up and Axel updated compare-locales to 2.7,

We also encountered two interesting bugs – Andre dove deep into ICU to fix `Intl.NumberFormat` breaking on roundings in Persian, and I had to disable some of our bidirectionality features in Fluent due to a bug in Windows API.

Notable changes [my work] [intl]:


With all that work in, we’re slowly settling down the new APIs and finalizing the cleanups and the bulk of work now is going directly into switching away from DTD and .properties to Fluent.

As Firefox 60 is getting ready for its stable release, we’re accelerating the migration of Preferences to Fluent hoping to accomplish it in 61 or 62 release. Once that is completed, we’ll evaluate the experience and make recommendations for the rest of Firefox.

Stay tuned!

Multilingual Gecko in 2017

The outline

In January 2017, we set the course to get a new localization framework named Fluent into Firefox.

Below is a story of the work performed on the Firefox engine – Gecko – over the last year to make Fluent in Firefox possible. This has been a collaborative effort involving a lot of people from different teams. It’s impossible to document all the work, so keep in mind that the following is just the story of the Gecko refactor, while many other critical pieces were being tackled outside of that range.

Also, the nature of the project does make the following blog post long, text heavy and light on pictures. I apologize for that and hope that the value of the content will offset this inconvenience and make it worth reading.

Continue reading “Multilingual Gecko in 2017”

One year with the Firefox OS L10n framework

For several years now, the Localization team at Mozilla has been working on a modern localization framework based on the following set of principles and architectural choices that we consider fundamental for the next generation multilingual UI’s.

  • Principle 1: Localizers should be in control of translations. The localization framework should be grammar-agnostic, whether it’s about grammatical cases, genders, or tenses.  Localizers should be able to use the entire expressive power of their language to author translations which create the best experience for the users.
  • Principle 2: Language fallback should be robust and graceful. When a translation is missing or broken the user should be presented with a translation into the next best language, given their preferences. There might be more than one fallback language.
  • Principle 3: Translations should be isolated and asymmetric if needed. The source language of the application should not define structure of the translations (e.g. the lack of pluralization in English should not make it impossible for other languages to use plurals in a given message).
  • Principle 4: The framework should embrace the Web. Localization should react to changes of the runtime environment (e.g. resizing of the app’s window, change of orientation, incrementation of a number of unread messages) and should add as little overhead for developers as possible.

Exactly one year ago, on April 8th 2014, Stas landed the initial rewrite of l10n.js – the localization framework used in Firefox OS. This set us on the path to enable the vision of a modern localization framework driven by the design principles outlined above.

Since then, we have a dedicated, two person team, working full time on advancing this vision and learning how to improve upon it in the process.

The full year of work has resulted in many important features being developed for the platform, including:

  • Language packs: Small packages that decouple language resources from the application allowing us to extend language coverage dynamically when users request it, even after the device has been already released on the market.
  • Pseudo-locales: Programmatically built language resources that emulate different languages allowing developers to test their applications for any multilingual problems before localizers have time to provide translations
  • Asynchronous l10n: Major shift away from using synchronous API’s for retrieving localizations from JavaScript. This results in clean, condition race free code that is easier to write and maintain. It also sets us on the path to enable runtime language fallback.
  • Security: While not traditionally big thing in localization, having an open runtime ecosystem of localizations requires us to make sure that translations cannot accidentally or maliciously impact our code and break it.
  • Error reporting: We’ve made major advancements to help developers and localizers find potential errors early. We reject malformed strings, report missing strings and duplicates, recover from exceptions in our code.

A couple weeks ago we have finalized the work scheduled for Firefox OS 2.2 and begun development for the next major release. The clean and reliable API has given us a good base to start implementing the remaining components of the vision behind L20n in this cycle.

For the current cycle we have scheduled:

  • DOM Overlays: Ability for a localizer to use HTML syntax in their translations and also provide whole localized DOM Fragments to be merged with developer provided skeletons via a secure algorithm. This increases the system’s security and empowers localizers to provide better translations.
  • L20n format: One of the last remaining pieces of the puzzle is the new file format that is designed to store localization data like multivariant strings, entities with values and attributes and variant selectors. This will allow us to start introducing new features to the system that are impossible using the current data storing formats.
  • Lightweight l10n contexts: Together with the whole platform, we want to make a heavier use of the concept of multiple small localization contexts to replace the single-context-per-app approach. It will improve performance and isolation resulting in easier maintenance.
  • API 3.0: Our current API still contains remaining pieces from the old, synchronous API that we’d like to remove. Together with lightweight contexts and on the path to WebAPI we will want to make sure to organize our events, methods and objects to fit the design of other W3C APIs.

With ever-growing understanding of the environment and how the web stack matures, we are also getting close to start extracting the core of our framework to offer for standardization, and that’s an exciting opportunity to fulfill the vision of both Firefox OS and L20n and bring the modern localization framework to the whole web, making it more multilingual and global.

Stas & Zibi

Crazy times for the localizers

March of 2011 is special for all of us. It’s the Firefox 4 release month and the whole project is operating in it’s “crazy” mode.

One particular plane on which it’s pretty visible is localization as we’re involved in almost every aspect of the release, every little detail, every step of the product release, every engagement campaign and of course in every bit of discussion about the state of the web that happens in our respective countries.

The story

As an organization that has glocalisation imprinted on the very fabric of its universe, we always feel responsible for making sure that everything we do is localizable, so that our communities around the world can translate and adjust everything that Mozilla project creates and use it to promote the Mozilla mission in their regions.

That’s easy to say, much harder to do. Especially as we grow, and the number of projects that we touch goes up. Firefox 4 release brings an unprecedented number of web projects that are part of our push to make bring the web to the next level.

Those projects vary from HTML5 demo sites, through Mozilla Developer Network Demo Studio, social web projects for twitter and facebook, web games to new Mozilla Foundation campaign – Join Mozilla. In total, I believe, we’re preparing 12 separate projects that we call “special projects” that are going to more or less closely accompany Fx4 release (not to mention Firefox 4 l10n itself, and our major projects like SUMO, MDN, AMO).

Each of the projects is very different and unique, uses different set of technologies and is embedded into different environment and yet we need to unify it for our localization community so that they don’t have to learn how to localize each of them separately.

Given the usual amount of challenges multiplied by 12 different setups it could have been a mess and it should have been impossible, but thanks to a great work done mostly by Staś Małolepszy, our team of l10n-drivers unified the experience as much as possible and localizers are right now doing an amazing job by bringing those special projects to every corner of the Earth, to billions of people.

Below is a summary of what happens there. Continue reading “Crazy times for the localizers”

Mini kampania z okazji wydania Firefoksa 4

Wraz z nadchodzącym wydaniem Firefoksa 4 Mozilla planuje małą kampanię promującą HTML5&Friends. szuka osób chętnych do pomocy w tłumaczeniu tej strony. Oferujemy dostęp do serwera testowego, niewielką ilość stringów do przetłumaczenia i sporą dawkę zasłużonego poczucia dobrze wykonanej roboty.

Zainteresowanych zapraszam do mailowania do mnie (zbraniecki at aviary dot pl) lub od razu do wejścia na Verbatim –  – zakładasz konto i dodajesz sugestie.

Uwaga, wysoka jakość tłumaczenia może skutkować niemoralnymi ofertami dalszej współpracy ze strony 😉

Using Common Pool – Tutorial

In this post I’m going to show you how to make your extension localizable via Common Pool and how it can be localized.

early adopter warning: It’s the first attempt, it’s a new concept, things may brake, things may be a little rough, stay calm. It’s not a replacement for a well known solutions yet, but it may be one day. Play with it, hack the service and the library, suggest improvements, report the issues, but stay calm whatever happens – there may be dragons.

Using the library

Since Common Pool is currently implemented in Addon SDK, I assume you’re using it for writing your extension. First thing is to add my localization module to your extension. It’s available at

Once you have it in your dependencies, you can just include it into your code.

Let’s take an example Hello World code:

var notifications = require("notifications");

var menuItem = contextMenu.Item({
  label: "Hello World",
  contentScript: 'on("click", function (node, data) {' +
    '  postMessage(node.src);' +
  onMessage: function (imgSrc) {
      title: "Hello World",
      text: "Welcome to the new, wonderful world!",

Now, in order to make it localizable, all we need to do is to add one require, and change the string initialization:

var notifications = require("notifications");
var _ = require("localization").get;

var menuItem = contextMenu.Item({
  label: _("Hello World"),
  contentScript: 'on("click", function (node, data) {' +
    '  postMessage(node.src);' +
  onMessage: function (imgSrc) {
      title: _("Hello World"),
      text: _("Welcome to the new, wonderful world!"),

Done! That’s what I meant when I wrote that Common Pool is asking as little attention from the developer as possible. If it reminds you of common patterns used by GetText, you’re right on spot. We like this notation, we reuse it.

With such code, you’re ready to go 🙂 As a developer, you can move along, do your work, deploy your extension on AMO and never ever bother about this again, while we will make sure that the localizers can localize, and users will receive the localized version dynamically. 🙂

Early adopter warning: For the time of the field testing of this library, we do not actually have any integration with AMO, so when you want your addon to be localized, you need to “cfx xpi” it, or if you’re using the builder just click on “Download” and upload the result xpi here: This is a temporary step

Using the web service

Once a developer uploaded his addon and his job is done, the camera shifts onto another hero – a localizer. This brave soul, can just go to, log in and start localizing. As a localizer, you can just select your locale and start localizing. You can filter by an extension and you can view how other locales localized it.

You can also select overrides and add one for a given jetpack or occurrence. When you finish localizing your jetpack, you’re done. We will spread your translation to all users of the jetpack who use your locale. If you later update your translation, all users will receive an update on the next day. Isn’t that awesome? 🙂

Early adopter warning: Until we fix bug 10, we refer to jetpacks by their unique ID, instead of the human readable name. It should be fixed soon 🙂

The beauty of early adopting

Well, that being said, you know you would not be happy to receive a complete, ready to use, bug free solution. That’s not what hackers life is about. In fact, we decided to blow quite a big hole through the middle of the road that’s called “First Run Experience“.

As we were working on the library, and the webservice, and how nicely they’re tied, and we enjoyed the experience of a user staying up to date with localization of his extensions, and being able to download an extension in en-US and then see it localized on the day when a localizer finishes localizing it… we actually did not solve a pretty important case. Here’s the issue:

When a user installs an extension it contains no translation resources bundled with it. On the first run, we asynchronously ask the localization web service for the entities used by the extension. So far so good.

The server asynchronously sends the user all the entities and they’re cached. From that moment every request for a string will work by providing a translation of the string. But what happens before we cache?

Well, for now, nothing. Bummer 🙁 Basically, until we cache the strings, all we have is an empty cache and the localization library will just return the english strings which is the default fall back.

The bug for that is reported, bug 619807, and we’re working on a solution for that. We may decide to bundle entities that we have available for the extension together with it at the time of download, but that will be just a workaround. An ideal solution would be to pull the localizations together with the addon (as a part of addon installation experience) and then check if there are any localization updates on a normal daily routine.

If you have any ideas, suggestions or code snippets that solves it – feel free to jump in and help! 🙂

Real Life Example

Take a look at this example extension, Translate Selection with l10n. Now, go to to see its localization and localize it live.

Next time you run this extension you should be able to see it working with localization (assuming your browser is using your locale and minus the First Run Experience bug).

That’s all. 🙂


Common Pool is a simple, yet very powerful solution. It solves a lot of issues that we’ve been struggling with since the beginning of times. It makes internationalization easier than ever, localization easier than ever and synchronization easier than ever. Bare in mind, it’s intended as a solution for simple strings, for developers who don’t have much time to think through their UI’s and plan it for localization, and for massive localization of tons of thousands of extensions.

We believe it nails it perfectly (with the few early adopters hiccups that we’re working on right now, and an unlimited number of issues that are awaiting to be discovered by you!), and we want to improve it until we feel it to be the right one.

At the same time, please, remember that for a more complex and sophisticated UI’s, the ones where a developer has to think about entities, and the localizer wants to fine tune each and every sentence – we’re going to work on a complementary, L20n based solution.

Now, it’s your turn to play with it, hack on it, give us feedback and help us evaluate the solution. We’re available at #l10n and #jetpack channels, and we use mozilla-labs-jetpack group that you can use. For bugs in the l10n web service, please, use its bitbucket issue reporting system. Thanks! 🙂

Jetpack Localization – Common Pool approach

Jetpack project is taking an aim at rewriting the extension system, learning from the past, responding to the feedback and thinking outside of the box. So when the Jetpack team approached L10n-Drivers and asked for support with building localization infrastructure for new addons, we decided to go the same way. 🙂

Common Pool

The first approach we’re presenting you is called Common Pool (yup, that’s not a very creative name) – it’s basic assumption is that most jetpacks are small applications with a minimal UI. And there are going to be tons of them. We already have tens of thousands of extensions, and Jetpack with its nice API and builder are going to make it much, much easier to write a good extension.

We also made some observations – There are common phrases, like “Cancel“, “Ignore“, “Log in” that are replicated over and over in almost every extension. It’s becoming hard to localize it, hard to keep it consistent and maintain the localization. On the other hand, extension authors are struggling to work with localizers as they need to iterate fast and  string freeze phases are adding a lot of cost to the release process.

For that reasons we decided to build on a basic concept of entities shared across extensions, stored on a server, and pulled by the jetpack itself.

Shared entities are powerful. For a localizer it means that we can show you the list of mostly used entities and if you localize 20 of them, you localized 50, 100 or maybe 500 extensions. You can update a translation and have it translated the same way across all extensions. You can localize only one extension and your work will probably benefit other extensions as well. You can localize today, and tomorrow all users will get updated localizations. How cool is that? 🙂

For a add-on developer it also means more good. You can pick entities that are already localized into many languages and have you extension Just Work. You don’t have to worry about releases and localization because if you only make thing localizable, users will be able to pull localizations as they appear. It means less to think about.

Now, there are of course some trade offs. One of them is that not all translations of a word should be equal. For that reason we propose a concept of an exception. Actually, two of them.

First one is a per-addon exception which allows localizers to localize a given string differently for this particular addon. It’s important when the addon uses some specific vocabulary.

Second one is a per-case exception which allows localizers to divert from the common translation for a specific use case. Say, the use case in jetpack X, file main.js line 43 should be translated differently.

Library and the Web Service

So, that’s the concept, and it has been implemented as a library. Yahoo! Now, for such a solution to work, we knew we need a partner in crime. We need a web service that will support Common Pool and allow for addon localization. We knew we will want to merge it with AMO and the Builder, so django was a preferred technology, and one other requirement was to be able to support a localization storage model that is substantially different from the industry standard.

Fortunately, at the time when we conceptualized Common Pool, our friends from Transifex were reworking their storage model away from gettext-specific to much more generic database driven one. Well then – good time to see how generic it can be, right?

Together with Transifex team, mainly Dimitris Glezos and Seraphim Mellos we’ve been working on the Common Pool web service that can back up the library and become a localization web tool. Today, the first iteration of the tool is alive at –! The source code is GPL and available at transifex-commonpool bitbucket repository and that’s where we’re collecting bugs as well.

What’s next?

Now we have to prove that the concept in the field. We’re challenging the status quo, so it’s on us to validate it. We’re going to reach out to extension authors and localizers alike to get support and testing coverage and if we that it works, we’ll include common pool library into Addon SDK 1.0. It’s up to you to help us evaluate it!

Then, it’ll be time to complement Common Pool with another architecture, that does not solve the simple cases that easily, and require much more active work from the developer, but will provide better level of details for both parties. The solution we want to offer is L20n.

Bundle pack ahead

Together, Common Pool and L20n should provide everything that localizers and developers may need pushing the limits and stretching what’s possible with the localization. We think of a Common Pool as a great generic, easy solution for most common strings, while L20n complements it with more sophisticated, more powerful approach for the developers whose extension reached the level when they want it to be localized with more attention and where the UI requirements are higher.

Let me know what you think – I know it’s early, I gave many talks about those technologies to our Mozilla communities, but I realize that it’s hard to evaluate without actual code in hand, so now is the time for feedback, feedback, and even more feedback 🙂

In the next blog post I’ll explain the steps to create your first common-pool driven extension.

Mozilla Summit 2010 – Localization 2.0 talk

Here come slides I used for Summit 2010 Localization 2.0 talk.

It was a very tough talk to give. Hard to grasp, hard to explain. I originally wanted to devote it exclusively to L20n, and make it as a form of tech talk, but eventually figured out it will not work and there’s much broader vision I need to explain. Thus a few hours before the talk I started rewriting it and end up with what you can find here.

Of course slides alone will tell you just a small part of the story, but it’s better than nothing. 😉

Thanks to all who participated in this! I know it eats a lot of brain cycles to process and it was already 4pm, but I hope you enjoyed it! 🙂

Pontoon – more details and 0.2 plans

Catching up from the blog post that introduced Pontoon, I’d like to present the project in a little more detail and the vision for a near future.


As stated previously, Pontoon 0.1 introduces two types of clients – web client and jetpack client. Both share majority of code base (Pontoon, jQuery, jQueryDomec and heavily modified editableText plugin), but because of the limitations of jetpack prototype I had to merge them into one JS file.

user experience

For user interface of Pontoon 0.2 I’d like to focus on a few key aspects that may make Pontoon a tool ready to use in work environment:

  • Introduce identification system that will sign contributions made by a given user. We may use Verbatim accounts for now.
  • Figure out a way to recognize the state of localization and allow for continuing localization.
  • Add ability to view/edit entity in a source view. (much like wordpress’s HTML view)
  • Fix some common cases like entity with a single anchor should treat the anchor as an entity, not its parent
  • Improve UI to look nicer and inform user on the outcome of his actions.

With those six improvements (including two top ones which are major), Pontoon will become, in my opinion, ready for real life use and we may offer it for one of the upcoming small websites for those brave souls who would like to test it.

code base

On the code base front, I will definitely want to clean up the sources, which will be easier when jetpack client get updated to new jetpack architecture. editableText plugin will require a cleanup and will be extended to allow source view, so it may make sense to fork it as a separate project, while other classes should be refactored and site-dependent code should be moved away from libraries. Also, CSS inclusions should be merged into libraries.

One thing I’d like to see soon is a third type of a client – one that may be embedded into any website so that the website author can turn on and off the “localization” mode.


Pontoon server will require adaptation to the client side changes, like ability to identify/authenticate users, and then two new things should be possible with the data received from a client. First, server should be able to add translations as suggestions in Verbatim. This way users will be able to suggest translations, leaving up to the registered localizers what to do with it. This workflow is an intended one for Pontoon.

Second thing, is that we should be able to generate a pure, localized HTML file out of a source one and translation list. This way we will be able to localize things like SUMO documents or plain HTML sites that do not use Gettext or even PHP.


Hooks will go through a major rewrite, because I want to test the concept of providing an external meta file with information like lists of entities for the client to read. Such meta information is the most natural way to also add ability to pick up localization from a midpoint. The issue here is that of course this will not help the client with localization of websites that do not use pontoon hooks.

It would be also great to add Django hooks, since so many websites this days are migrating to django, and it would also allow us to test how cross-linguistic the hook system is.


This is my plan for a short term, and I’ll try to get as many of the features described above into Pontoon 0.2 as possible. I prepared a wiki article describing the plan, and I’m open to feedback, ideas, and of course, contribution. 🙂