Loca what?


I've been meaning to get a blog started for some time, but XP SP2 has been taking a lot of time lately. Now that's pretty much done though, so allow me to introduce myself.

I'm Jesper. The Swedish Windows localizer. I work in Windows International. We're the ones who bring Windows (and more) to the world. No, really.

The localization team in Windows International has in-house localizers for over 20 languages. We manage and perform localization of all kinds of projects, from small utilities such as SMS Sender to slightly larger projects such as this, and for an increasing number of languages. The most recent release is of course Windows XP SP2, which is on its way to Windows Update right now.

What's cool about this job and this group is that we get to see a wide range of products and we get broad exposure to each product. In Windows, I translate everything from event log messages that help you trouble shoot AD replication, to aeronautical terms hidden somewhere in DirectX. We encounter loads of different products and while we don't become experts on each, over time we get a pretty good overview over the range of products Microsoft has.

Another great thing about the job is the combination of linguistics and technology. It's not always easy to translate software; you're often working without context, with new terminology and with ambiguous source text. Therefore, we spend a lot of time researching terminology, discussing translations with terminologists "back home", playing around with software to make sure our translations make sense in context and bugging developers to correct spelling here and changing a term there.

Fun stuff indeed.

This will have to do as an introduction. In upcoming posts I'll talk more about how our products get localized and new stuff coming out of this group. If you have any suggestions for topics to cover - anything related to localization or localizability - feel free to give me a shout.

Comments (11)
  1. rajaram says:

    gr8 topic .. awaiting ur new posts on localization, try introducing code snippets iff possible

  2. rajaram — thanks for your comment. Code snippets can be done, anything in particular you want code on? I deal mostly in localization, not so much in developing localizable software 🙂 That said, I have a lot of fun writing tools to help us work smarter & raise quality.

  3. Uwe Keim says:

    So what are the tools you are using during translation?

    All in-house, self-developed, or things like WinTransRC or way more professional tools?

  4. Uwe, we mainly use in-house tools for Windows localization. I think this is because back when we started localizing software, there were no 3rd party apps that were good enough (this was before my time, so I’m just guessing here).

    The main tool we use is developed within the Windows International group, and has a lot of good features for e.g. sizing dialog boxes, glossary management, recycling translations. For managed code, the tool simply wroks as a wrapper around winres.exe. Quite handy; even though the projects are slowly changing, we can use the same tool for almost all files we localize (there are a handful of exceptions, mostly legacy files).

    We also have a bunch of smaller tools & automation solutions in place, for managing glossaries, finding "bad" translations etc.

  5. Uwe Keim says:

    Thanks for your fast feedback, Jesper 🙂

    I really would like Microsoft to release all/some of the tools to the public, maybe as "unsupported tools".

    Now THAT would be really really cool!

  6. Uwe – I haven’t heard any plans to release the main tool publicly (that’s why I don’t even refer to it by name :), but winres.exe is out there for managed code.

    Also, check back in a couple of days — my next blog will be about a solution we’re working on that’ll probably include some tools. This one is directly targeted at ISVs, to help them bring localized software to the people faster & cheaper.

  7. Eusebio Rufian-Zilbermann says:

    It’s good to see a blog about localization.

    As suggested topics, how about localization processes (e.g., dividing in-house vs outside work, what makes a good lockit, spot linguistic checks vs. thorough linguistic reviews, resource freezes). Also it would be very interesting if you could post on localization advocacy (even better if you have some ‘hard’ data e.g., what percentage of people and companies really use localized products? is it true that for server products many companies in non-english-speaking countries still prefer the US version? how much do sales increase after localization?)

  8. Eusebio – great suggestions! These are all issues we debate & constantly work on improving internally.

    I don’t have the data you ask about handy, but I’ll do my best to find it and share in upcoming blogs. What I do know though is that a significant portion of our revenue comes from outside of the US and localization is becoming increasingly important for the company.

  9. Eusebio – I found something. http://msdn.microsoft.com/msdnmag/issues/04/03/Bugslayer/ says "Microsoft makes over 60 percent of its revenue from outside the U.S.".

    This doesn’t mean that 60% of all software sold is localized, I suspect that a fair portion of the sales is English versions abroad. We’ve increased the language count, but there are still a lot of languages left to cover…

  10. Actually, other figures that I had were similar. Nadine Kano, in the foreword to the 2nd edition of Developing International Software mentions that in 1989 sales of localized products had reached 50 percent of Microsoft’s revenues and recent figures (2003) are closer to 60 percent.

    Considering that these numbers refer to localized products, I’d say this is slightly better than saying that 60% of the total revenue is coming from outside the US.

    An argument where I’ve been involved a couple of times when doing globalization and localization advocacy is that for certain software (e.g., server apps) it would sell the same regardless of whether it is localized or not. Of course I don’t subscribe to this opinion but I’ve never found any statistics that prove or disprove the point (maybe nobody has really researched it yet?) and it just ends up in a "speculative opinion contest".

    Thank you very much! 🙂

  11. To make it harder to track who’s running what version, we have MUIs and LIPs as well… (I’ll write something about this next week, the different levels of localization)

    About non-localized server, I hear the same. Apparently, the general consensus seems to be that sys admins are better at English than your average user. I don’t buy that. I think it’s the opposite – we localize server products in fewer languages than client products; we make documentation such as TechNet, MSDN, KB Articles available in very few languages… thus, sys admins are more likely to pick English OS.

    From this follows that the right problem to address is to make it easier to use localized servers. Stuff is happening in this space. The research group in Microsoft spent time on buildin some really cool natural language parsers and put them to use in machine translating a huge amount of KB articles into some languages (e.g. Spanish). The machine translated articles weren’t perfect, but as with our LIP projects, they were a whole lot better than nothing.

    There’s still work ongoing in trying to find a way to provide support and documentation in more languages. I hope over time that we’ll be able to release servers in more languages and that using a localized server will be the natural choice for any sysadmin.

    Oh, and another argument I hear now and again is that security fixes are released quicker or are more reliable on English os. Not true. These days we always simship fixes for all languages, and all languages receive the same testing (far as I know — feel free to disprove me if you can:). The only exception is Service Packs which get staggered releases, but they’re different. 🙂

Comments are closed.

Skip to main content