On Jun 17, 2005, at 6/17/05 7:59 AM, Bullard, Claude L (Len) wrote:
> Why avoid attributes? That is superstitious behavior.
Whether or not individuals share the theology, that's the choice the
TC has made, repeatedly. And I don't think the issue here is large
enough to justify us going back and re-plowing that old and bloody
ground.
> If you name it for it's origin, you will add hundreds of elements.
> If you codelist it, you only have to include that and maintain it.
See, this is where trying to compromise gets a guy. ;-)
Look, I remain persuaded that we can and should specify a single
CRS. I was merely trying to address a concern about the spec alone
possibly not being enough to clarify which CRS that is. I was NOT
trying to open the door to forcing implementers to support hundreds
of CRSs.
(At most they should have to deal with two... WGS84 and their
preferred local CRS. I suspect the vast majority will continue to
get along just fine with WGS84 alone.)
Again, I haven't heard any actual implementer report any of this as a
problem... nor have I heard any technical reason why it should one.
This whole topic seems to be based on a hypothetical concern for some
third parties' sensitivities. Seems like we may be chasing ghosts here.
- Art
