Ben Laenen wrote: > It's really the cleanest option and allows for additional tags like a > name, and it allows everything to be more clearly defined,
I don't see why that would be the case. "For every way with tag X, Y applies" isn't any less clearly defined than "For every way in a relation with tag X, Y applies". What you probably mean is that we wouldn't need to define new tags with a set of implications, but could instead create a relation that directly contains all tags that would otherwise be implied. I'd still prefer a standardized tag for use in relations, though. Adding all relevant tags (the proposal page lists several suggestions here) to the relation for every single city is very error-prone and doesn't allow easy changes of those sets. > and allows > for a method which is much more flexible and allows for extended > properties. This is correct, using relations for restrictions does allow for more flexible tagging, especially with stuff like proper conditional restrictions. However, these advantages aren't limited to what we are talking about here (standardized sets of default restrictions). Instead, they are even _more_ useful for explicit signage (whether zonal or otherwise). The standardized restriction sets can easily be expressed with a single tag, simply because there aren't many of them. So while there are reasons for using relations as a means of expressing restrictions on ways, I don't see why this relates to this proposal. The "shortcut" tags for commonly implied sets of restrictions could be used in relation-based mapping as well. Tobias Knerr _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
