On Wed, 23 Apr 2008, Andy Robinson (blackadder) wrote: > Because it was difficult for the "layman freeform tagger contributor" to > decide what the root "class" should be, for instance is it class=waterway or > class=river.
I think I'd be inclined to try and make things a bit hiararchical - e.g. "waterway:river" or "waterway/river". But really this sounds like a documentation issue more than anything - if there is documentation providing reasonable guidance, and a good set of existing classes all following the same "standard", this seems like a good way of doing things. > Just going back to namespaces (trying not to nag you), the use of namespaces > is very useful in certain contexts, where true separation within a dataset > is desirable. My main concern with a lack of namespaces is that you can end up with a single tag key meaning very different things depending on the context it is used in, and the context is not always obvious. This makes interpretting the tag, looking it up in the wiki, etc. a problem. And if you attempt to unify the meanings of these tags so that they are not so ambiguous when you don't know the context, you end up having to coordinate far too many bits of the project. > Using a string of namespaces in front of each and every tag (the key getting > longer and longer as the data gets more complex) doesn't really in my view > give the same flexibility that I think is needed for OSM. To some extent, this is about presentation - if the editors presented this string of namespaces in a nicer way, rather than just a big string of namespaces then it would suddenly become much nicer to work with. - Steve xmpp:[EMAIL PROTECTED] sip:[EMAIL PROTECTED] http://www.nexusuk.org/ Servatis a periculum, servatis a maleficum - Whisper, Evanescence _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk