Google is getting old. The search is limited.
"What? What do you mean by limited?!" Limited by language. The new search engine should find what you are looking for in all common languages. If I'm looking for "head of state" it should yield the french, german, spanish, dutch, tagalog, mandarin, japanese, korean, canadian (heh), and portugese versions of the search as well. Of course, to do this well would take some intelligence (which Google does not lack), but I think a rough version of this is past due.
Does anyone know of a search engine that does this?
Subscribe to:
Post Comments (Atom)
6 comments:
The first step is getting it to recognize synonymns in english. I believe you can do that by putting a ~ in front of the word. But yes, language equivalence would be nice. Alternatively, allow translated versions of the page to be searched. (I believe that would be a lot easier.)
au contraire (as the french might spell better). I think it would be easier for the system to find the translated word rather than have the full translated pages be catalogued (crawled)... maybe that is what you meant. heh, you up for a bit of a challenge? (c: using 'babel' you could probably write an equivalent web portal. upstart google?
Hey Luke,
My humans prefer clusty.com to the big G. It may not translate but it sorts the sites.
Woof!
Freda
hey freda, thanks for the tip! I'm quite surprised i've never heard of clusty.com before. It seems very clean and well ordered. i'll start testing more in the coming days. (c:
Luke: I know it seems counterintuitive, but if it translated your own queries it would have to do it on the fly every time. If it translated the pages, it would only have to do it once, and then it could store it. It's a question of which is cheaper, storage or computation. (Hint: ususally storage. But not always.)
Of course, that only makes sense with a single language; once you throw multiple languages into the equation then your way makes more sense. (Three points for Luke.)
Anyway, I would do it myself, but in a week and a half Why the lucky stiff will come out with a new Ruby program that does all the same stuff in 172 lines of code making judicious use of web services and harnessing the collective intelligence powers of t3h web 2.9! (With tagging, I hope you realize. I mean, it should be obvious by now.)
Freda: you have humans? Awesome.
They have 2.9 now? Man, I'm behind the times...
Post a Comment