Quick comments
Posted: Fri Jul 14, 2006 3:38 am
- AA seems to have settled down to something sensible.
- Don't notice much difference in PO defence yet; maybe a little bit less inclined to attack. Formations with defend orders can still charge off into the wild blue yonder away from their objectives if there's nothing much in front of them. Formation order semantics and modelling have always been screwy since TOAW I, IMO, and could do with reworking from scratch; not just tinkering.
- Turn-end processing time reduced by maybe half on average for the medium/big thing I'm working on. But variable - some turns can take as long as before. That's what I would expect from tinkering with the algs. Still think that changing reconstitution placement to something simpler would be best way of getting further (& desirable) speed ups, rather than alg-tinkering.
Actually the biggest speed improvement I notice so far is in AI movement - presumably uses the same path-finding?
Further speed-ups I think most important for testing; probably only marginal importance for playing, given the improvements in this patch. But to the extent Elmer vs Elmer is now a good first test for a scenario, you really want the thing to be fast enough to run multiple tests for most changes you make which might effect balance or event logic - maybe 10 or more runs. Not very feasible if a run takes 1 hour+, which my biggish-but-not-huge scenario does at the moment.
- Did the withdraw-army bug get fixed? haven't tested that yet, or the various event editor/force editor bugs.
- Don't notice much difference in PO defence yet; maybe a little bit less inclined to attack. Formations with defend orders can still charge off into the wild blue yonder away from their objectives if there's nothing much in front of them. Formation order semantics and modelling have always been screwy since TOAW I, IMO, and could do with reworking from scratch; not just tinkering.
- Turn-end processing time reduced by maybe half on average for the medium/big thing I'm working on. But variable - some turns can take as long as before. That's what I would expect from tinkering with the algs. Still think that changing reconstitution placement to something simpler would be best way of getting further (& desirable) speed ups, rather than alg-tinkering.
Actually the biggest speed improvement I notice so far is in AI movement - presumably uses the same path-finding?
Further speed-ups I think most important for testing; probably only marginal importance for playing, given the improvements in this patch. But to the extent Elmer vs Elmer is now a good first test for a scenario, you really want the thing to be fast enough to run multiple tests for most changes you make which might effect balance or event logic - maybe 10 or more runs. Not very feasible if a run takes 1 hour+, which my biggish-but-not-huge scenario does at the moment.
- Did the withdraw-army bug get fixed? haven't tested that yet, or the various event editor/force editor bugs.