Thu, Apr 03, 2014 at 12:07:42PM +0200, Janko Mihelić wrote: > 2014-04-03 11:12 GMT+02:00 Martin Koppenhoefer <dieterdre...@gmail.com>: > > > > > FWIW, it is not true, we would "save" 1 way or 2, but the amount of nodes > > would remain the same, because with the new proposal the waterway would get > > an extra node which it hasn't otherwise. The 1 way saved is on the other > > hand loss of information as pointed out before. > > > > I went super-geeky, and made a quick estimation of xml size: > > http://pastebin.com/ZDXPv8fK > > I don't know if the database sizes are proportional to xml, but there's 674 > letters in the brunnell version, and 1306 letters in the bridge version of > the xml. So more than double :) Waterway only refers the same node.
thanks. I suspect the difference would be more substantial if the bridge would carry several hiking routes or you want to put a waterway into a culvert which happens to share way with 3 admin boundaries? Also, it is frequent practice to "anchor" the other way which would add other (up to) two completely useless nodes. More than database size the other issues bug me much more: * we pretend a level of precision which is not there and knowingly enter junk data into the database. * the resulting structure is more difficult to cleanup and maintain than a single node Richard _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
