W3C home > Mailing lists > Public > semantic-web@w3.org > March 2011

Re: Why skolemization?

From: Sandro Hawke <sandro@w3.org>
Date: Wed, 30 Mar 2011 09:16:56 -0400
To: Richard Cyganiak <richard@cyganiak.de>
Cc: Steve Harris <steve.harris@garlik.com>, semantic-web@w3.org
Message-ID: <1301491016.2139.758.camel@waldron>
On Wed, 2011-03-30 at 17:59 +0530, Richard Cyganiak wrote:
> On 27 Mar 2011, at 05:53, Sandro Hawke wrote:
> > http://garlik.com/=rdfgensym=/6135eb5943eaed2
> 
> More appropriate would be to use RFC 5785 [1]:
> 
> http://garlik.com/.well-known/gensym#6135eb5943eaed2

Yes!  Good idea.   That solves the architectural problem of "squatting"
URI space.

It looks like non-Hash URIs are also possible, like:

        http://example.com/.well-known/gensym/6135eb5943eaed2

since RFC 5785 says:

        [The spec] MAY also contain additional information, such as the
        syntax of additional path components, query strings and/or
        fragment
        identifiers to be appended to the well-known URI
        
So, if you had many different services on a host each generating their
own Skolems, they could be allocated their own subspaces like:

        http://example.com/.well-known/gensym/sandro-service/77211
        http://example.com/.well-known/gensym/steve-shop/d97c0666-5ace-11e0-b3aa-00216a3e966a
        ...

The webserver can handle the delegation to the service with a rewrite
rule.   That should work fine.

This wouldn't formally apply to tag: URIs, but it should be fine in
practice, and perhaps that could be declared in a revised tag URI spec. 
A formally correct: I could define "tag:hawke.org,2011:gensym:uuid:" as
a tag prefix for uuid-generated gensyms.   Since tags aren't
dereferenceable, it doesn't really matter that it's hawke.org in there.
Also, tag:hawke.org,2011:gensym:tag:" as the prefix for gensym tags,
like tag:hawke.org,2011:gensym:tag:garlik.com,2011:bnode:6135eb5943eaed2
The point being that with his prefix you know it's an RDF gensym,
which ... might mean something (like that you should try harder to merge
it, or something).

Now we just need someone to write a gensym spec.  The hard part being
explaining what these things are good for, when you should use them, and
what you can do with them...   :-)

   -- Sandro

> [1] http://tools.ietf.org/html/rfc5785
> 
Received on Wednesday, 30 March 2011 13:17:06 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:26 UTC