Re: Summary of the QName to URI Mapping Problem

>Pat Hayes asked:
> > This seems to me to be the basic issue: who has the authority to say
> > they are distinct?
>
>In other words, who gets to decide what a name identifies?

Not exactly. There are two issues here: when two names are the same 
name, and what exactly it is that a name is supposed to identify. The 
latter authority  has to be given to the 'author', I agree. But the 
former question seems more like a matter of an agreement between the 
author and the intended audience, to use the same lexical/syntactic 
conventions.

> The generally
>accepted answer on the web is that if the name has an authority component
>(i.e. a host name), then the (current) owner gets to decide.  You can of
>course ignore the owner's edict and make up your own mind.  (If the owner is
>spouting contradictions, that might be a good idea!)  If multiple
>authorities make claims about a name, you have to decide which one to trust.
>
>I wouldn't trust the RDF spec with regards to identity, though.  If I write
>up a nice XML document with non-identical QNames from a namespace whose
>domain I own, the XML Namespaces spec allows me to claim that they identify
>different things.  Along comes RDF,

But wait: if you are writing open-form XML, then RDF shouldn't even 
be able to *see* it. RDF only reads RDF, even when it is reading 
XML-RDF. So if you want to be understood by an RDF engine, better use 
the RDF rules; and if you don't, then by all means use some other 
rules. Seems fair to me.

> concatenates the pairs, and concludes
>that the QNames identify only one thing.  It contradicts the document's
>author, whose claim was perfectly valid based on the standard definition of
>QNames.  This is a bad thing.

It would be if they had made that claim in RDF.  But (apparently) 
they didn't, so why should it even matter what RDF might make of it?

> > If they really were distinct, that is. Or, you could take the
> > position that the use of the mapping shows that they couldn't have
> > been distinct.
>
>That indeed seems to be the position of people defending the current
>mapping.  It's not a problem as long as everybody agrees to it.  However,
>since the mapping is not part of the XML Namespaces spec, and not everybody
>uses the RDF interpretation, I don't think this is something we can impose
>on the rest of the world.  And if RDF chooses to use this mapping within its
>own space, then its definition of QName is no longer compatible with
>everybody else's.

Again, I take your point, but I think you mis-state the case. Nobody 
is saying that arbitrary XML Qname usage is expected to follow RDF 
rules. But if someone sets out to use RDF, then why would it be 
unreasonable to expect them to abide by the published RDF 
conventions? Isn't that the point of publishing a spec, to "impose" 
the conventions on anyone who chooses to use the system being 
specified?

Pat Hayes

---------------------------------------------------------------------
(650)859 6569 w
(650)494 3973 h (until September)
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Friday, 17 August 2001 17:36:39 UTC