On 07.06.2012 17:33, sly (sylvain letuffe) wrote: > A recent proposal (and change after that) on the wiki has been made, which > roughly sums up to : "relations type=multipolygon's members should only be > closed ways, not sums of ways making closed rings, unless the way is too big > that it would be refused by the API" [...] > I think that's an important change to one of the mostly used relation type > (multipolygon) which should be a little more discussed before validation.
Let me first point out that these edits don't affect the definition of the multipolygon relation per se. No relation would become "invalid", so this doesn't break anything. It is about which of the alternative mapping styles is preferable when either would work from a technical point of view. My opinion here is that multipolygon relations should not be used unnecessarily. So if a closed way is possible (no holes, "small" area), then it should be used instead of a multipolygon. Other cases are less clear, but I generally think that multipolygons with non-closed outer ways are being over-used beyond their original purpose and make dealing with our data more complex than it needs to be. The original reason for extending the definition of the multipolygon relation to allow non-closed outers were very large areas, and imo their use should mostly be limited to those cases. Tobias _______________________________________________ Tagging mailing list Tagging@openstreetmap.org http://lists.openstreetmap.org/listinfo/tagging
