Re: web proper names

On Tue, 21 Sep 2004 16:59:21 +0100, "Chris Purcell" <cjp39@cam.ac.uk>
said:

> > It is here that the need for this distinction seems clearest. Witness:
> >
> > <http://www.paris.org/Monuments/Eiffel> ex:resultOfDereferencing
> > <http://www.paris.org/Monuments/Eiffel> .
> >
> > at which point meaning disappears up its own anus. It isn't clear that
> > there is anything in the world of the web that can be the result of
> > dereferencing itself.
> 
> One could accuse you of excess pedantry here. ex:resultOfDereferencing 

My word. I'm shocked! You want excess pedantry? Pedantry is by
definition excessive!

Ahem. Sorry. Couldn't resist. I am, er, known for pedantry. 

> as a predicate can mean "the subject is the result of shoving the URI 
> of the object through any known, applicable de-referencing mechanism". 
> This is quite clear, and will confuse neither generic RDF engine nor 
> any software which understands the predicate.
> 
> Many existing RDF applications have the implicit assumption:
> 
>      for all x that can be de-referenced, <x ex:resultOfDerefencing x> 
> holds

How is my software to know what any given x can be dereferenced, without
trying it? And if, upon trying it, my software finds that it can be
dereferenced, I still can't distinguish between the case where the
result is that which is indicated by the URI, and the case where the
result describes to a human reader what it is that is indicated by the
URI.

> and that this assumption violates the statement in the RDF Primer:
> 
>      RDF uses URIrefs only to identify things, while browsers also use 
> URIrefs to retrieve things

It also violates the right of the URI coiner to decide what their URI
indicates, does it not?

> However, we don't need to conform to this requirement in the definition 
> of ex:resultOfDereferencing.

Why not (besides that the Primer is non-normative)?

One man's pedantry is another's clarity? In this case it seems to me
that the options are 

a) to create properties which instruct software to treat their object
URIs in a non-standard way (that is, to treat them as (URI qua retrieval
path)s), introducing special cases where the object URI indicates,
instead of what the URI indicates, the URI itself, or 

b) to use the already available mechanism for distinguishing (URI qua
symbol)s from (URI qua retrieval path)s: use typed literals for the
latter. 

There's even the anyURI datatype there ready and waiting for case b. 

Cheers,
Hamish

Received on Tuesday, 21 September 2004 16:37:09 UTC