Millers rpl2 improved the execution time of
[:(] <@~./.~&, ((>./ ([EMAIL PROTECTED],: <.//.) <./) rpl ])^:_) |:
by some 20%.
If you only need the size of the components, then
[: #/.~@, [: ((>./ ([EMAIL PROTECTED],: <.//.) <./) rpl2 ])^:_ |:
gains another 20%.
I especially like the part
>./ <.//. <./
R.E. Boss
> -----Oorspronkelijk bericht-----
> Van: [EMAIL PROTECTED] [mailto:programming-
> [EMAIL PROTECTED] Namens John Randall
> Verzonden: dinsdag 3 juli 2007 18:29
> Aan: Programming forum
> Onderwerp: RE: [Jprogramming] FW: Components in graphs
>
> R.E. Boss wrote:
> > I made a merge of comp and mcsg
> >
> I am very pleased to hear this.
>
> > However I realized that it is rather inappropriate to take into account
> > points which are not mentioned in the edges.
>
> This too: I have been using a technique to get around it that is much
> cruder.
> >
> > But recently I discovered another approach which appears to be much
> faster
> > even:
>
> I will try this out shortly. I have two concrete problems where my holdup
> is calculating the size of components in a large number of graphs. I do
> not need the exact vertices, but I eliminate cases based on component
> sizes. Since your previous method did not work fast enough for me, I have
> been eliminating some cases using gross ad hoc criteria, but it would be
> much more elegant to just calculate the component sizes.
>
> Best wishes,
>
> John
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
