[WEB4LIB] Re: Even stranger AltaVista result

[Charles P. Hobbs]

On Wed, 17 Nov 1999, Tara Calishain wrote:

> 
> >One more question - how does AltaVista decide which language?
> >Is there some sort of automatic analysis and comparison to
> >frequently used words on the site from some sort of word list?
> >This was asked from one of the participants - at the time I just
> >wanted to step down and laugh hysterically.
> 
> I asked AltaVista this last April. Bear in mind that they might have changed
> their method, but last April they said that when their spiders index a page, it
> sets a language tag on it as soon as it encounters a sufficient number of words
> in a given language.

It might default to English, until a sufficient number of foreign language
words have been read. Often, I've seen multi-lingual pages, or pages in
languages that AV doesn't know (for example, Arabic), reported as English.
. .