On Thu, Nov 17, 2005 at 07:50:13PM +0100, sejourne_kevin wrote:
> Oliver Bandel a écrit :
> >Any hints here?
> >(Maybe using OCaml, but the imperative features of it would help,
> > if the functional features would be too slow?)
> >
> >Any hint on algorithms/datastructures for this would be fine...
> >
> 
> In my work we use "Lucene" (apache.org). Lucene is in java but have been 
> re-coded in few other langage (c++,python...), the spec of the infex 
> format is availble for free (online).
> 
> I think the real problem for 10^x (x > 6) docs is the size of the index, 
> not really the speed of the answer(not for lucene (fast and small)).

The size of the index as well as creating time of it.
If it's not updateble and must be created new for every
update, than this needs to much time.

If updates are cheap, then thats the right thing.

Ciao,
   Oliver