Re: namespace usage as assertions

>From: "Manola,Frank A." <fmanola@mitre.org>

>To wit:  isn't it the case at the moment that, since the URIs that
>identify namespaces may actually point to nothing, the most you can say
>about using such an identifier is that it's an assertion (by the user)
>that there is a namespace identified by the URI, and that the namespace
>contains the specified element name(s)?  

No.  If I have <a:x xmlns:a="http://a.b.com" /> then all I am asserting is
that the globally unique identifier of this element type is a two-part
object, the first part being "http://a.b.com", the second "x".  

Now, if I'm a good citizen, and want other people to be able to do things
with my "x" elements, I ought to publish a prose description of what 
this element is all about, an open-sourced reference implementation of
software that does something useful with it, and a schema (in descending
order of importance IMHO but reasonable people may differ on that :))

But it would be perfectly OK to say nothing, never have the intent of
putting anything at http://a.b.com, and still use this namespace without
violating the letter or spirit of the namespace REC.

What *would* violate the spirit of the namespace rec?  The creation
and publishing of a fragment such as that above there being any software
which knows about this element and can do something with it.  Read the
first paragraph of the REC.

I also think the use of a relative URI ref totally violates the spirit
of the namespace REC, but I don't think we have consensus on that.

It turns out that the choice of URI-ref syntax for namespace names not
only gets you pretty-strong global uniqueness, it also opens the door
to a more semantically-driven web in the future, based on using the URI
semantics of the namespace name.  We don't know how yet... but it's 
interesting.  -Tim

Received on Wednesday, 21 June 2000 14:11:08 UTC