On 30 August 2010 14:18, Richard Welty <[email protected]> wrote:
> On 8/30/10 9:06 AM, M∡rtin Koppenhoefer wrote: > >> +1, but site-relations might still be useful in the context of power >> generators. There are situations where the single objects do not >> overlap but are side a side, for example you might have 3 generators >> with 3 chimneys and want to model which chimney is connected to which >> generator. >> > i'd lean towards site relations being useful because i think that > the computational complexity of doing lots of polygon intersections > is being underestimated. yes, for small bounding boxes it's ok, > but consider if you needed to do it on a larger scale, it'd make > certain tasks completely unreasonable (i'm not sure what those > tasks might be yet, haven't thought about it.) > Based on this discussion, it seems that the best advice to put on my proposal for power generators is: - use site relations where the power=generator objects don't obviously overlap with the buildings they relate to, particularly where you are dealing with a cluster of nearby objects - for simple cases like a wind turbine on a roof, it is helpful but not necessary to use a site relation to specify the relationship between the generator and the building Regards, Tom -- http://tom.acrewoods.net http://twitter.com/tom_chance
_______________________________________________ Tagging mailing list [email protected] http://lists.openstreetmap.org/listinfo/tagging
