Re: Moving on (was Re: URIs quack like a duck)

David Carlisle wrote:

> > Then how do you reconcile relative URIs with that requirement?
> They are allowed but give you a namespace name which isn't globally
> unique as do several absolute URI
> data:,david
> mailto:david
> file:///dev/null
>
> If the creator of the namespace desides he can live with this
> lack of uniqueness then so be it. (but I wouldn't object to
> the namespace rec using stronger language to discourage the practice)

I agree.  I think there's a larger principle at play here: that a spec
should be usable by people other than the ones who were involved in
writing it.  That means that when there's useful context, it should be
included.  That's what the non-normative stuff is about, I presume.  And
a non-normative appendix on recommended practices and the reasons
they're recommended would be extremely useful.

> This lack of uniqueness is no reason at all to change the namespace
> spec so completely that it creates some kind of new unstable
> XML file that can be valid as it stands but if you use cat
> to send the contents to standard out, the result isn't even namespace
> conformant.

Agreed.

Paul Abrahams

Received on Thursday, 1 June 2000 12:28:08 UTC