[OSM-talk] namespaces

[Dave Stubbs]

>  I don't really see reason for such a heated discussion. We are
>  already using namespaces. Most implementations of namespaces somehow
>  support the concept of a "default" or "root" namespace, which is
>  where, logically, all those tags currently reside that are not
>  qualified with a namespace (i.e. 95% of tags). We can move tags out
>  of the default namespace into their own namespaces if and when it
>  makes sense; new tags can be created with their own namespaces if it
>  makes sense, old tags can remain where they are unless there's a
>  pressing practical problem.


The reason for any heat is that when presented with a tag name
piste:lift:occupancy, my basic reaction is why would you do that? It
doesn't do anything for you other than make the tag name longer. Now
if the response is, "well, I think it looks prettier", then that's
hard to argue with, but the argument has frequently been that this
somehow lets us do magic. Which of course it doesn't. And it annoys me
that people do things on a misunderstanding.

Your points 1), 2) & 3) are all correct, although I'd further add to
2) that we're probably talking about extensive generic attributes --
ie: not ones that apply to specific object classes. Your examples of
architecture or civil engineering details are spot on. These are the
kinds of things you might expect an editor to "fold"-away, even if it
didn't understand the specific tags. A perfect example of where this
could happen already is the tiger: namespace used for all the import
meta data.
Normal information about a lift, on a lift object, definitely comes
under category 3).

Dave