What can we expect of YAPF?
I saw it pass on SVN ( http://svn.openttd.org/cgi-bin/trac.cgi/changeset/4182 ) and wondered what it's functions are?
I haven't been here for a while but i've checked almost every day the progress of OpenTTD by this URL: http://svn.openttd.org/cgi-bin/trac.cgi/timeline
It sounds pretty useles to do and it is
YAPF?
Moderator: OpenTTD Developers
- webfreakz.nl
- Director
- Posts: 627
- Joined: 11 Aug 2005 08:22
- Location: Localhost, 127.0.0.1, [The Netherlands: South Holland-> Westland]
- Contact:
YAPF?
# Programming is like sex, one mistake and you have to support it for the rest of your life. (Michael Sinz)
- LordOfThePigs
- Route Supervisor
- Posts: 435
- Joined: 01 Jul 2004 10:28
- Location: Jura/Switzerland
webfreakz.nl: As you can read from SVN log message:
[YAPF] - New branch for 'Yet Another PathFinder' based on trunk r4181. Attempt to create fast and flexible pathfinder.
- Later the new PBS should be implemented on top of it.
What more do you want?
Old PBS had 2 major issues:
- dependency on NPF, which is very slow
- it didn't work very well (many bugs)
So if I want to make new PBS, I need yet another pathfinder to build PBS on top of it.
Yes, it would be nice to have pathfinder agnostic PBS, but I think that PBS could be made much better as pathfinder feature - it needs extra support from pathfinder:
1) PF must record the path if PBS should reserve tracks on the best path
2) PF must respect the PBS reserved tracks and find another way if possible
Unfortunately NTP doesn't record the path.
There are also many complains about NPF performance when used for other transport types (water, road). NPF is well designed but not very optimized for performance. NTP is much faster but cryptic and therefore not ready to be easily extended.
I would like to make another PF, that will combine good (flexible, understandable) design and high performance. Later I want to build new PBS on top of it.
[YAPF] - New branch for 'Yet Another PathFinder' based on trunk r4181. Attempt to create fast and flexible pathfinder.
- Later the new PBS should be implemented on top of it.
What more do you want?
Old PBS had 2 major issues:
- dependency on NPF, which is very slow
- it didn't work very well (many bugs)
So if I want to make new PBS, I need yet another pathfinder to build PBS on top of it.
Yes, it would be nice to have pathfinder agnostic PBS, but I think that PBS could be made much better as pathfinder feature - it needs extra support from pathfinder:
1) PF must record the path if PBS should reserve tracks on the best path
2) PF must respect the PBS reserved tracks and find another way if possible
Unfortunately NTP doesn't record the path.
There are also many complains about NPF performance when used for other transport types (water, road). NPF is well designed but not very optimized for performance. NTP is much faster but cryptic and therefore not ready to be easily extended.
I would like to make another PF, that will combine good (flexible, understandable) design and high performance. Later I want to build new PBS on top of it.
Who is online
Users browsing this forum: No registered users and 42 guests