W3C home > Mailing lists > Public > www-rdf-interest@w3.org > July 2004

Re: InverseFunctional properties are the new URI?

From: Dan Brickley <danbri@w3.org>
Date: Fri, 30 Jul 2004 05:56:55 -0400
To: Damian Steer <damste@hplb.hpl.hp.com>
Cc: Sandro Hawke <sandro@w3.org>, www-rdf-interest@w3.org
Message-ID: <20040730095649.GC17970@homer.w3.org>

* Damian Steer <damste@hplb.hpl.hp.com> [2004-07-30 10:26+0100]
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Sandro Hawke wrote:
> |>Actually, I think I'll disagree with myself before anyone else does.
> |>Taking Dan's point, the ordering could well be IFP > no URI/IFP > URI
> |>because the URI is in no way a property of the described object whereas
> |>all other properties are.
> |
> |
> | Why isn't something's URI an IFP property of the thing?   TimBL calls
> | that property log:uri, I think.   For a while, I generalized it
> | slightly to u:uname [1].
> |
> |      -- sandro
> |
> | [1] http://www.w3.org/2001/12/uname/
> 
> Hmm. Why not use rdf:resource and rdf:about (being samePropertyAs)?
> 
> Backwards compatible, and parsers will just need to de-bless those
> attributes. For example:
> 
> <rdf:Description newrdf:about="http://ex.com/1">
> ~  <a:prop newrdf:resource="http://ex.com/2"/>

If newrdf:resource is unblessed, ie. 'just a property' this is same as 

<a:prop>
  <rdf:Description>
    <newrdf:resource>http://ex.com/2</newrdf:resource>
  </rdf:Description>
</a:prop>

...which isn't quite what you want, I think. You'd also run into
back-compatibility problems since properties can repeat, while rdf:about
can't. Also if rdf:resource is a property, it would appear in a
different syntactic role, ie. as an element. And rdf:about could appear
there too. Thinking about it, it'd be total chaos :)

Nice try though! Maybe if all next-gen serializers were constrained to
write out a constrained (and possibly lossy) idiom...? Hmm even then I
think the property elements couldn't work in old and new worlds at same
time (eg. are these literal-valued or resource-valued properties? do
they take the string form of a URI? if so, property elements get written
differently).

Or maybe I misunderstand the above example. Perhaps if a:prop was
owl:sameAs it Might Just Work? If it wasn't for the back-compatibility
aspect, this would be a simplification of the syntax. With
back-compatiblity in mind, it looks somewhat more scary...

cheers,

Dan


> produces what you want. s/newrdf/rdf/.
> 
> Damian
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
> 
> iD8DBQFBChQxAyLCB+mTtykRApzZAJ0bYLuIF0h/S7U1+CT97vwtPZ6FLACg2vcx
> CLEM2dyc5Qm36+WyaDj82gw=
> =llMw
> -----END PGP SIGNATURE-----
Received on Friday, 30 July 2004 05:57:39 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:58 UTC