Practically I don't think it will work, as it requires much more data to be processed in the preprocessing - I therefore agree with Martin.
But I think an intermediate solution should and could work: If all entrances to the area are not accessible (e.g. gates, lift-gates and such), adding access=no/private to them should work. Personally I would consider routers to be buggy when they ignore barriers tagged on nodes of the way, while I would accept them not to do geometrical calculations between areas and ways. Keep in mind though, that e.g. for forests often access is only allowed on the highways that cross it. Access=no at the forest area would - following your argumentation - imply access=no on any highway crossing it, which will fail in most situations, because nobody explicitly states access=yes to the highway in these cases. regards Peter Am 20.03.2014 06:39, schrieb bulwersator: > Is it reasonable to expect that well tagged road contains all access > tags necessary to check whatever it is accessible? > > In other words - is it OK to tag area like proving ground with > access=no, without tagging roads on this area with proper access tags? > Should cases like this be reported by validators? Or maybe routing > engines should be expected to process also areas that share no nodes > with processed ways? > > http://wiki.openstreetmap.org/wiki/Key:access#Nodes.2C_ways_and_areas > basically says "do not remove correct access tags", is not answering > my question (it is about removing tags, not adding new ones. Also, it > is failing to consider that it is possible to ensure adding missing > tags by using quality assurance tools). > > http://wiki.openstreetmap.org/wiki/Computing_access_restrictions is not > mentioning areas, maybe because roads should have full set of relevant > tags, maybe because it is incomplete. > > Note: this is not purely theoretical, I encountered it during planning > development of JOSM plugin. > > In my opinion all relevant access tags should be on way and its nodes, > otherwise it is unclear whatever road inherits access data from area. > > Moreover it would make processing data overly complex, but this one > may be easily countered that requiring specific access tags would make > editing tedious. > > > > > > _______________________________________________ > Tagging mailing list > Tagging@openstreetmap.org > https://lists.openstreetmap.org/listinfo/tagging > _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging