Log for on 16th March 2014:
Times are UTC Toggle Colours
07:00:19  *** Supercheese has quit IRC
09:46:28  *** Alberth has joined
09:46:28  *** ChanServ sets mode: +v Alberth
10:06:40  <fonsinchen> for FS#5941
10:12:37  <Alberth> this uses 3/2 ?    In the issue, frosch speaks about using 4/3.
10:15:07  <Alberth> I'd prefer to have fewer ways to measure things rather than more, unless that creates problems
10:16:45  <fonsinchen> This uses the pre-existing DistanceMaxPlusManhattan which is already used in the ship path finder.
10:17:23  <Alberth> ah, nice
10:17:26  <fonsinchen> 4/3 would be a new invention and it would give me more headaches as it's < sqrt(2)
10:19:40  <Alberth> Wrt. the sqrt(2) thingie. Would it make sense to use the same  approximation as in Vehicle::GetAdvanceSpeed resp. GetAdvanceDistance?  That is sqrt(2) = 4/3   <-- this states another opinion on that, but if 3/2 is also used already, that is fine to me
10:21:22  <fonsinchen> If the approximation is less than sqrt(2) then the errors accumulate to shorter and shorter distances when you add them. This works against the stopover penalty in the MCF solver.
10:21:44  <fonsinchen> If the approximation is more than sqrt(2) it just generates higher stopover penalty, which is fine.
10:24:20  <Alberth> sounds like a convincing argument not to use 4/3
11:23:26  *** frosch123 has joined
11:23:26  *** ChanServ sets mode: +v frosch123
17:39:37  *** Knogle has joined
18:20:45  *** Supercheese has joined
18:20:45  *** ChanServ sets mode: +v Supercheese
18:45:15  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26403 || Logs: || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
21:56:36  *** frosch123 has quit IRC
22:01:59  *** Alberth has left

Powered by YARRSTE version: svn-trunk