I saw this flier pined on a bulletin board outside of the Computer Science Department at MSU the other day:
At first I thought this might be a Computer Science class on translation algorithms. That would actually be kinda cool and interesting. I never had any experience with building this kind of tools, and I would love to learn more about it. If I had to build one, I would probably look into neural networks, or use TF-IDF like approaches to do statistical analysis on tons of sample data – or at least that would be the starting point for further research. I’m pretty sure there is some impressive shit being done in that field these days.
But then I noticed how the course was labeled Linguistics 450. Translation – it’s not a and not Computer Science course. It is a non-technical course about shit that Linguistics people probably have no idea about. I really can’t figure out what is the educational value of this course. Without actually looking at the algorithms and methodologies involved in translation this course has as much value as a Discovery Chanel documentary on the same topic.
I just don’t like when non-technical people try to teach others about technology – because, oh boy, do they get it wrong. When we do research that overlaps with another field (for example bio-infomatics) we are expected to have strong background and extensive research done in that area. But when other people do research that involves technology, they just tend to make all kinds of crazy assumptions without a shade of understanding of the real issues at hand, and then just wing it.
If anyone is taking this class, let me know if it’s crap or not. I’m suspecting it will be, but I would love to be pleasantly surprised by it.
[tags]translation, bablefish, engrish, translator, linguistics[/tags]
Is teh computer revolution passing YOU by?
Yeah, I know. WTF?
I’m going to go out on a limb and presume you know nothing about modern Linguistics….
No, I can’t say I have done any extensive research on modern linguistics. I’m not belittling it or anything. I think it is a very interesting area of study. For example, J. R. R. Tolkien was a linguist and I was always impressed by his ability to construct fictional languages with fluidity, grammar, and etymologies that actually made sense (eg. Quenya).
But I’m sure a CS class on designing translation software would require at least a basic understanding of linguistics. I will go out on a limb and say that it is probably easier to teach computer scientists basic linguistics rules, than it is to teach linguists programming, text parsing, and modern textual analysis paradigms…
I might be wrong though, and I would love for someone to explain to me how this class would work, what topics it would cover, and what would students be expected to get from it.