Re: removing constraints on 'resource' [024-identity]

On Fri, May 28, 2004 at 07:30:03AM -0700, Larry Masinter wrote:
> Are you saying that you think the scheme definition *doesn't*
> define the range of things identified by URIs using that scheme?

Well, more like "shouldn't" than "doesn't", since some scheme
registrations clearly do restrict the range.  But yes.

> I'm wondering why you think we have different URI schemes, if the
> resource-identifier to resource-identified mapping isn't defined by
> the scheme. 

I believe we have different URI schemes because we have different
information spaces, and a URI needs a means to indicate which
information space its information comes from.   But IMO, that's
orthogonal to the type of resource that the information represents.

> > But it seems to me that removing that
> > part wouldn't change your message substantially.  Any objections?
> 
> It leaves out some interesting and useful information.
> Maybe you could explain something that is harmed
> by the assertion in question?

It suggests that a proliferation of URI schemes is desirable, because
there are obviously an unbounded number of types of resource that can
be identified.  I consider that harmful due to the large cost of
deploying new URI schemes, i.e. that until a new scheme is well
deployed, possibly useful information is unavailable to many (like
those darned itms:// Apple URIs I keep clicking on).

> So while I agree that the original you're correcting is awkward,
> I'm not happy with the amendment. Frankly, I'd just as soon leave it.

Ok, no biggie.

Mark.
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca

Received on Friday, 28 May 2004 15:20:07 UTC