At 2011-08-21 10:57, Henk Hoff wrote:
For every rule we can find exceptions.
In this case, I will guess the exceptions (shared routes) are less than 5% of the ways.
The basic idea behind the decision-tree was: use the most important / most logical route for the way-ref tag.
If you know the "important one", make it the first value in the series.
Putting every single route-label in the ref-tag is not a good idea.
Why? I guess that 95% have only one, 4% have two, and the remaining 1% might have more (I seem to remember seeing 4 in the midwest somewhere). Remember we're only talking about the road routes themselves. Bike routes, etc. go in their own tags.
If you want to identify a whole route, use a relation. Based on the relations (a way is part of) a routing engine could then identify under which other route numbers this road is also known by.
As someone pointed out, once you put them in a relation, the tags on the ways become duplicative. While this is generally bad database design, it's also true that many consumers don't deal with relations, and so we need the duplication and the problems that go with it.
-- Alan Mintz <alan_mintz+...@earthlink.net> _______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-us
