Archive

Archive for the ‘Operating System’ Category

One forward and two to the side

25/06/2012 2 comments

The debate about digital technology and localization and internationalization has probably raged in one form or other ever since someone invented the first program. Mind, for me personally it goes back to that ill-fated moment when ASCII was born with some bright spark arguing that no one would ever need more than those few letters that English has. My first computing headaches were around ASCII – how do I do an /ɣ/ and what the heck was %73£ when someone typed it at the other end?

Much has happened since and I’ve moved from phonology to software translation big time but I still can’t quite decide whether we’re in a better place now or not when it comes to small languages. Those technicalities (like ASCII vs Unicode) aside, the field has indeed opened up, in particular when it comes to open source software. There’s nothing but laziness that stops a language from having at least an office suite (LibreOffice), a browser (Firefox or Opera), an email client and calendar (Thunderbird and Lightning), a media player (VLC), a wiki (MediWiki), a spellchecker, a forum package (phpBB) and blogging software (WordPress.org and .com) – satisfying a fair chunk of your average user. For the really tough there’s Linux in all its scary glory of course. Ignoring the height of the bar when it comes to actually localizing some of them, that’s not the whole story though.

At least in digitized countries, a significant chunk of our work and social lives have shifted onto various digital platforms. Desktops, laptops, smartphones, tablets… you name it. Hardly a year goes by without some innovation hitting the headlines. And the tech savy (overwhelmingly the young) have become real digital nomads. Yesterday’s app is so passé today and today’s market leader mobile phone OS may be tomorrow’s digital roadkill (anyone remember Symbian?). It’s a bewildering, fluid place.

It’s a place we can’t ignore. Whether we like it or not, virtually anyone under the age of 25 has a smartphone, from rocky outcrops in the Western Ocean like Barra to the mountains of Gipuzkoa, the deserts of Arizona and the steaming hills of Papua New Guinea. Ok, maybe not Papua New Guinea yet though it wouldn’t surprise me. The more of a space we can carve out for out languages and cultures, the better because sadly the old maxim of “Use it or lose it” – or however your language puts that – is true.

So we must compete somehow, at least at some base level. But I increasingly feel that without a small but dedicated full time team, this will become harder and harder unless there’s some magic on the way that I haven’t heard about. Let me give you an example. Predictive texting goes back to the 1970s, believe it or not but not wanting to be too depressive about it, it probably did not make huge inroads into our lives before the year 2000 or so when it really took off on phones. Back then, you had those languages which your manufacturer deemed appropriate, maybe a dozen or so if you were lucky. We’re now in 2012 and I’m waiting with bated breath for the first release of Irish, Scottish Gaelic and Manx on Adaptxt which, after much searching, I discovered last year. Finally an open source predictive texting project open to any language. Yay! Ok, so it only works on Android… I can live with that, looking at the Android market share. It would be good if iPhones also supported 3rd party entry methods but they don’t and I’m getting to the cheesed off stage with Apple’s approach to non-billion-speaker-languages anyway.

But I digress. There we are, happily preparing the tool which will finally take Scots Gaelic and Manx out of the letter-by-letter age (Irish has had Téacs since 2008 but I’m not sure how alive the project is) when Apple starts pushing Siri (that voice recognition thing on iPhones which, by the way, only works if your accent resembles that of the Queen and or Charleton Heston). I bet my bottom dollar that before long, every major mobile phone manufacturer will be running something similar.

Here, I gnash my teeth. Predictive texting is reasonably easy to do as long as you have a framework you can feed your data into. For example a spellchecker. But it’s taken around a decade for such a framework to grow out of the cyber community. Speech recognition is a harder. A lot harder. I have no idea how long it will take for languages such as Gaelic to take that hurdle and even less so of how many of this planet’s 6,000 languages will manage to do so. And that makes it all a little frustrating.

I don’t know what the answer is, right now, I just feel it would be nice if stuff slowed down a bit. Honestly, how much technological innovation do we need in 12 months? Or rather, how many false summits can we and our languages keep pace with?

Advertisements

Detect locale – manna from heaven or hellspawn?

15/12/2011 6 comments

It seems like a good idea, doesn’t it? Web 2.0 and all that, increasingly intelligent software taking the task of selecting your language when you visit a page.

Based on? Aye right, there’s a catch. Based on one of two things – the language of your browser or the preferred languages you can set in most browsers. Those of you who are normal end users have probably already spotted the problem. That’s all those of you (the majority) who went “I can do that?”. Yes, you can, but for the vast majority of people, what I shall call the “Install and Hope” group, that’s both news and several steps too technical. And before you tut at that degree of inability to tweak software, that probably includes your mum and dad, your aunts, uncles, grannies and gramfers. They’re not stupid people on the whole.

While actually the more intelligent choice because you can select from a relatively wide range of languages, this still limits you immensely. There are some 6,000 languages on the planet and even multilingual Firefox only offers maybe 200 or so in the language dropdown. And many of those are languages like Chilean Spanish, European Spanish, Argentine Spanish – what the codemasters call locales. So what are you supposed to do if you’re a speaker of one of those 5,800 NOT on the list?

But here’s the thing that drives me and many other speakers outside the club of the 25 biggest languages insane. Most sites these days take the lazy approach and just base it on the language of your browser (most websites), or worse, your operating system (Linux and most mobile phones). In the words of Julia Roberts “mistake, big mistake”. Why? Because there are far fewer browser localizations than languages. And even if there is a browser localization in your language, that doesn’t mean everyone is using it.

Let’s take your average family on the planet which is – believe it or not – bi- or multilingual. That usually means that between the parents and kids at least two languages are used. Sometimes even more. But not everyone usually speaks both languages but, and this is where it gets tricky, they will often share a computer. Say we have a bilingual English-Kurdish family with a Kurdish-speaking father and a wife and kids who only speak English. The default language on the computer will most likely be English because in most cases, you can’t install the same browser more than once and few offer you an easy way of switching the language. So the browser is in English. But for the sake of argument, say the father wants to blog in Kurdish using the Kurdish version of WordPress. He goes to the main page and looks for a list of localizations. Tough luck, there isn’t one, because WordPress.org relies on your browser language settings. So he downloads the English version even though there IS a Kurdish version, it’s just not obvious because you have to go to http://ku.wordpress.org/

It could be worse – they could be using Ubuntu. True, it’s become more user friendly but who designed that insane bit of forcing the locale? Let’s say Azo, the father, wants to install some other software only he will use in Kurdish. What are you supposed to do in Ubuntu? Right, you go to the Software Centre. Only problem is, someone again figured that tying the language of any software you download to the OS language is a bright idea. Not for those of us who aren’t monolingual. And no, suggesting that Azo goes to the address bar, types about:config, types matchOS and toggles to false and then selects ku in general.useragent.locale is NOT a solution.

This dance gets even more insane. Let’s say Azo would like to use the Kurdish version of a mobile browser on his phone at least, since he can’t get Kurdish WordPress. Unfortunately, he uses an Android phone. Meaning? Well, while there IS a localization in Kurdish, the language of his Android phone is English, so the phone assumes this person could only ever wish to use stuff in English and forces every installation to English. End of.

Proprietary or OpenSource, same difference, Linux or Android, Windows or Mac, language selection is getting more and more difficult these days in spite of a legion of volunteers who strive to localize stuff into their languages. For free, usually.

So the question is, why, if there are localizations available for stuff, do you guys make it SO hard to get them? Isn’t that like baking a beautiful cake and then hiding it in the basement, assuming that everyone knows that’s where you hide the cakes?