View Full Version : Other resources on path solving
ErosOlmi
07-10-2007, 02:01
Hi all,
I've found the following library that comes in source code and binary format. It has a nice demo that can inspire some new ideas:
http://pathlib.hildebrand.cz/pathlib.html
Maybe interesting for experiments or ideas.
Ciao
Eros
Good find, Eros. Looks very interesting. The documentation in in Czech, but the author says the code has English comments.
Michael Hartlef is working on a path module, I'm looking forward to seeing what he develops!
Petr Schreiber
07-10-2007, 09:49
Hi,
if you will find any problem with fact it is Czech,
I can translate parts you are interested in for you :)
But as we have A-Star now and DJISKRA is comming from Mike, I think we have / will have some things to examine already
Bye,
Petr
Michael Hartlef
07-10-2007, 10:32
Hi folks,
my pathmodule is basically done but compared to randalls on a 50x50 grid, it is wicked slow. They say that the algo is slower then A* but I didn't thought it would be that much. So I'm searching for a mistake I might have made or for an allready made implementation in one of the languages I have to see it this will have the same results.
Michael
ErosOlmi
07-10-2007, 10:39
Michael,
are you doing something on script or all the hard job is inside compiled module?
As you may have seen, Randall developed module in such a way all the data is defined inside the script and module receive pointers to the data. In this way all calculation is done in compiled code and script is responsible mainly to maintain user interface and drawing side.
Ciao
Eros
Michael Hartlef
07-10-2007, 11:00
Eros, you should know me better. Hard work is done inside the module.
ErosOlmi
07-10-2007, 11:29
Yes, no doubts about your capacity Mike.
The reason I sometimes ask stupid questions is because I'm always worried about some errors inside thinBasic Core or something I can improve.
Sorry for my question.
Eros
Michael Hartlef
07-10-2007, 15:07
No problem Eros, I should have put a smily in my message. I solved one bottle neck, now it is faster by a factor of 10. Still, on a 50x50 grid, empty with no walls, A* is faster.