Software Internationalization, SIMS Style

Internationalization of software is incredibly challenging. Consider this Wikipedia sandbox page in Arabic, which is a right-to-left (RTL) language:

This is a companion discussion topic for the original blog entry at:

You should listen to Sigur Ros’s album from 2002, if you haven’t heard it. It doesn’t really have a title (it’s known by “()” because of the art on its cover), and the album artwork, cd, booklet, etc, feature no words. All of the “lyrics” in all of the songs are complete jibberish. The whole thing lacks language.

I haven’t heard any of these Simlish recordings, so I don’t know how they work, but when listening to (), your brain seems to want to recognize words very badly, and so you kind of map the sounds into words roughly and they come out weird. It makes you feel kind of like your internal speech-to-text unit crashed.

Reminds me of the loituma girl.

Wow! What a great post. Just so long as I don’t have to know Simlish to get past the capcha rapcha.

Hi! Very interesting subject.

I think you’re mostly wrong about Simlish and human conventions across languages. Simlish works well for westerners because it’s based on English structure and experssion.

There are many cultures which sound and behave completely different than English. For their members the tone in which Simlish is spoken and even the physical gestures by the Sims are unintelligible. They just don’t make sense.

Of course, that red octagon with a white hand is still going to be a problem for people in nations that whine about the colour red or open palms. And anyone who take their cow worship seriously (to the degree of letting the poor things starve to death because, well, feeding them can be expensive) are liable to whine about the hamburgers.

So, anyone remember what happened to Esperanto?

How did that “new language that isn’t any existing language but has similarities to many other languages” work out?

Oh, yeah. It turns out that hardly anyone was actually interested.

And tooltips were invented because it turns out that icons are really amazingly hard to use if you want to convey anything that isn’t already known by every 2 yr old on the planet. (i.e. “stop” signs are easy because everyone knows about them. “antidisestabilsmentariainism” is a really hard thing to convey in an icon, because very few people know what it is.)

Simlish is just another foreign language (one that can’t ever be learned), and I got the exact same impression from the songs as I used to get from french, japanese, russian, indian, hebrew (etc) songs. Especially french, and I suspect this is because french and english are so similar in construction and intonation - indians or chinese might get the basics of the emotion, but without the same linguistic pattern nuances would be lost. Try to figure out what emotions hebrew is conveying, for instance - it isn’t quite what your gut would first tell you.

Honestly, the emotional content of music - which is primarily used for tranfer of visceral emotion, not storytelling, except in folk music - doesn’t seem to have anything to do with translation and localization. There’s just nothing to localize if you’re not conveying any deeper meaning, and for that reason people can play some import games no problem. The speech is a total red herring, the side channel information is what’s being used to process it all.

Likewise, a red octagon is a stop sign in most countries whether it has english or not (most of the world does use english stop signs…), so there’s nothing to localize. Swedish signs are even more iconic, but some are incomprehensible without foreknowledge. In general, the history of traffic signs has evolved into reasonably descriptive icons that can be learned and then applied anywhere - if your point was how well that applies to software, well, it was quite hard to pick out of the post. I totally believe that fascilitating fast understanding with pictures and sound is very important, which is why I love programs with icons in menus, and websites that try to balance gobs of text with mystery meat - standardization is the key.

If you’re just writing about cool stuff in a stream-of-consciousness format, well, cool. =p Sorry, this was a little long for a comment.

Side note, awesome sign vandalization.

(Typically, someone said it far more succinctly than I could. .)

It is an interesting read. I would consider the idea of Simlish as a very novel solution to a complicated problem though.

I think that what’s often forgotten beyond language, cultural symbols, social structures, or other surface elements like colours, is that culture not only affects our perceptions, but also thinking processes and manners of interacting. It’s ridiculous for us to assume that even if we were to ‘solve’ the problem of localization, all cultures would value and use the same ‘tools’ (like websites such as Wikipedia) as those of us do in western societies.

I think the problem of localization will be addressed only when all societies are equally empowered to develop their own tools, or at least be informed enough to be able to determine what they locally need.

i think you mean “Antidisestablishmentarianism” sorry to be a pedant

Having watched those music videos in Simish, I cannot help but remember a webcomic from ten years ago:

It’s a bit too early for the idea of completely giving up on verbal and written communication, even though both fail me occasionally. :slight_smile:

P.S. Funny, but the excerpt from “Levan Polkka” that is used in loituma.swf, while sounding vaguely finnish, actually is the only gibberish verse in the song, meant to imitate an instrumental solo part, which isn’t otherwise possible a-capella. :slight_smile:

I enjoyed reading this article very much!

Nice, man!

It took me several years to learn “properly” english… and to tell the truh, I have been deeply enjoying the gibberish on english sang tracks… And I continue doing, most of the times human voice is just another wonderful instrument. Being able to grasp what are they singing about is the cream on top or the key point to like rubbish music with fantastic lyrics…

Right. We should avoid I18n altogether by using Ido. English is a mess.

Simlish works well for westerners because it’s based on English structure and experssion.

Urig, I have no doubt The Sims is biased towards western culture (or at least romance languages), but can you provide specific examples of where you think the Sims breaks down for non-westerners?

And yes, the the hamburger sign is probably not a great choice for India.

I don’t thing the reataurants with the hamburger signs are a good thing in India… They should have chosen a more neutral sign for a really international sign… A hard thing to do… An apple, perhaps?

I kept hearing that Lily Allen song as though it were in French! I wonder if that’s just me or whether it actually sounds like French?

The whole thing reminds me a little of what I felt like when I first heard about Racter. Of course it became apparent that it was a hoax, however the idea is still fascinating. Just like meaning without language (Simlish), Racter was (supposedly) the manifestation of the idea of language without meaning. With nothing but syntactically correct junk, we are left to make up the meaning based on our own preconceptions. Would different people derive the same meaning from text output from a real version of Racter? What if you translated it?

Language is a strange tool. Many argue that language, if used skillfully, can convey any human concept or emotion. However, I have often wondered whether or not language simply boxes concepts in, limiting them to preconceptions and association, rather than allowing the free transmission of pure thoughts and feelings. Perhaps language is a form of lossy compression of these things, which, no matter how complex and powerful our brains are, can never be fully recreated? Or maybe our current language is simply a flawed way of expressing ourselves and we may someday invent a lossless compression of our ideas.

Urig points out that Simlish works for western listeners. That’s not exactly correct. It works for non-tonal listeners.

For example, Japanese and Korean are non-tonal, so these sounds are somewhat familiar (except for the "R"s and "L"s).

It’d be interesting to create a tonal Simlish language. Maybe hire some Vietnamese and Chinese people to create it.

Even then, those who speak vowel harmonic languages are left out, like Magyar (Hungarian). Can’t please everybody. :wink:

Anyone remember Pingu? All the dialogues were improvised by one person. Admittedly, elements of German, Italian, French and English seep in - but it’s another example of a very expressive ‘nonsense’ language.

Jeff And yes, the the hamburger sign is probably not a great choice for India.

Obviously, that’s a Gardenburger.

This reminds me of the phenomenon of people enjoying opera more when it’s sung in a foreign language – even if not the original one – than if it’s translated to their own. When you can understand what the singers are saying, that interpretation of language kind of preoccupies your brain and prevents you from just hearing the musical emotion.

And this talk of localization vs. translation reminds me of the translation fascination of Douglas R. Hofstadter (of iGoedel, Escher, Bach: An Eternal Golden Braid/i fame). He wrote an entire book about translating one (not-very-long) french poem, and just how much to translate.