xml-dev - Re: Dereferencing a namespace URI {WAS: Success factors for the WebandSe

["Sean B. Palmer" <sean@mysterylights.com>]

> Again, what about URNs?  What about proxies?  What about DNS bandits?

In my own defence (heh, heh):-
Lets get this sorted out. TimBL at XML200 suggested (as is my
interpretation) that URI's should be:-
     a. dereferencable
     b. use the HTTP scheme
This means that they should be independant of TCP/IP and all of the other
technologies we have now: they should be permanent for the next 1000 years
or so.
Statement:- 1. XML Schemas allows the derefenrencing of URIs. 2. RDF needs
this to survive.
Problems: 1. What scheme should the Schema being referenced take on? This
is a "duh" answer. namespaces are a fundamental part of XML, so use XML
Schemas, or whatever current W3C recommendation there is for it (boy, am I
going to get flamed for that comment...I don't care though!) 2. Namepaces
aren't definitive. This is the *real* problem.
I suggest that if the fundamental architecture of HTTP and URIs were more
stable, we wouldn't have any problems, but as it is we do.
So, in summary, the W3C recommends that namespaces be used for
dereferncing, even if this is not evident in the namespaces specification.
Further to that, although it may be useful for the URI to point to a Schema
of some sort, this is probably a cracked architectural principle.
Overall (in relation to the Semantic Web) I don't think it matters how many
people agree/disagree/moan/flame/+ me on these principles, because that
doesn't alter the fact that it's usable/unusable/+ What we *must* attepmpt
to do is solve the fundamental problems with the system: namespaces aren't
definitive etc., and if that cannot be solved, then the whole architecture
of the Web is built on quicksand. If was can solve them, we have to look
for some new problems :-)

Kindest Regards,
Sean B. Palmer
http://infomesh.net/sbp/
http://www.w3.org/WAI/ [ERT/GL/PF]
"Perhaps, but let's not get bogged down in semantics."
   - Homer J. Simpson, BABF07.