On Mon, May 17, 2010 at 7:07 PM, Michael Minutillo
<[email protected]> wrote:
> This is how it works and I'd bet that the IntelliSense index is based on
> some kind of trie (or a http://en.wikipedia.org/wiki/Radix_tree) built from
> this data. You can still do a "contains" search with a trie but you'd have
> to index every possible substring (which could become a time/space concern):
> i.e.

Exactly. And I'd suggest you actually don't do that and implement some
sort of strategy to decide the "most likely" "contains" typings; and
then index them in a slightly different way.

It would be really fun to take a crack at it, actually.


> --
> Michael M. Minutillo
> Indiscriminate Information Sponge
> Blog: http://wolfbyte-net.blogspot.com

-- 
silky

  http://www.programmingbranch.com/

Reply via email to