Re: Advanced data structures in Glib ?

On Tuesday 12 March 2002 17:13, Owen Taylor wrote:
> If RBTrees were really vastly better than AVL trees, then we should
> just replace GTree with them..

I guess it depends if the little speed improvement associated with the 
single pass rebalancing in RBTree is sufficient to switch the GTree 
implementation. I think it is, but of course it's not my call.

> From my perspective, I think requests of new data types in GLib
> should be phrased as:
>  - We need a data type in GLib that does XXX. It might be implemented
>    internally as a foo-bang whazzit.

I can't speak for everyone, but from my perspective "we need a data 
type in GLib that associates a key with some value, *without* 
removing/replacing the value when an insertion is performed with an 
already existing key. It might be implemented with either a hashed 
table that deals with collisions, or similarly to STL with some kind of 
a tree (RB being the most common)"

> (And of course, willingness to implement would help as well :-)

I'm very much willing to do that.

> Regards,
>                                         Owen

Zack Rusin

Does this .sig make my butt look big?

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]