Hi,
What is it exactly we are talking about? It appears to me that
multilingual search can be interpreted in at least two ways:
--1-- putting together a database that comes up with a translation in
several languages of words you type in to search for, so that the
search engine itself knows what to look for... At least you should know
the languages you select for the search, but for the rest this seems
straightforward enough to implement as long as you restrict yourself to
latin characters.
--2-- putting together a unicode-based search engine that enables you to
type in any language in any script you would like and perform a search
on it...
The Logos browser demonstrates the idea, but this pretty thing is not
for use on the web. Reason is that the underlying datastructure used by
the Logos browser is very much standardized, and specially designed for
and aimed on supporting multiple languages. This is something we can't
say for the web. Perhaps XML will bring us more standardization, but it
still implies a thorough revision of existing indexes and indexing
methods. Seems like a real challenge and lots of work to be done.
Aernold.
> -----Oorspronkelijk bericht-----
> Van: Limin Zhang [SMTP:limin@sylvia.harvard.edu]
> Verzonden: vrijdag 14 augustus 1998 20:14
> Aan: Unicode List
> Onderwerp: multilingual search
>
> Hi, another question: what is the current development and future
> trends of multilingual search on the Web (i.e. multilingual
> information
> retrieval). Thanks
>
> Limin
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:40 EDT