skolem

On Tue, 2011-04-19 at 11:15 +0100, Andy Seaborne wrote:
> 
> On 19/04/11 10:19, Steve Harris wrote:
> > On 2011-04-18, at 21:37, David Wood wrote:
> >
> > > Please see:
> >>
> >> http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2011.04.20
> >
> > I think it's worth appending Sandro's description of the .well-known
> > prefix to the skolemisation proposal.
> >
> > I left it outto save space, and because it was fresh in most people's minds anyway.
> 
> You had to be there!
> 
> >
> > - Steve
> >
> 
> Trying to be clear about the details for myself, I gathered together the 
> emails and minutes:
> 
> Final proposal from the minutes 14/April:
> 
> ----------------
> 
> PROPOSED: If systems are going to reveal Skolemized bnodes, without 
> doing damage to the graph, they MUST use a fresh URI (per bnode) and 

Note that the language about "fresh URI" was added late in the
discussion and I don't really like it since its redundant -- that's what
Skolem terms are (fresh), by definition.

> SHOULD follow the form
>    http[s]://[domain]/.well-known/genid/[locally-uniq-id][#]
> or
>    tag:[domain],[year]/.well-known/genid/[locally-unique-id]
> (or, someday, genid:...). Such IRIs are considered more disposable.
> 
> "genid" to be reg'd with IETF.
> 
> ----------------
> 
> Observations:
> 
> 1/ Why is
>    http://host.com/.well-known/genid#1816
> excluded?

An error.

> 2/ The authority is not optional in http URIs (RFC 2616 sec 3.2.2).

I was typing this during the meeting and ended up overloading a made-up
template syntax and BNF syntax.    I'm still not sure the right way to
say this in ASCII.   The [s] meant the "s" was optional; the [domain]
meant that a domain name went there.  (really an authority, yes.)

> A skolemized bNode might look like:
> 
> http://host.com/.well-known/genid/cb48cfae-6a5d-11e0-ba3a-00270e03441
> 
>    (that's using a version 1 UUID)
> 
> tag: can be used when machines don't know their own external name (e.g. 
> on a 10.* network).
> 
> 
> Subsequent email (key parts):
> 
> On 16/04/11 15:46, Steve Harris wrote:
> 
>  > My suspicion is that the only way forward would be some text along
>  > the lines of: [with apologies for any abuse of terminology]
>  >
>  > Systems wishing to skolemise bNodes, and expose those skolem
>  > constants to external systems (e.g. in query results) SHOULD mint
>  > fresh a "fresh" (globally unique) URI for each bNode.
>  >
>  > All systems performing skolemisation SHOULD do so in a way that they
>  > can recognise the constants once skolemised, and map back to the
>  > source bNodes where possible.
>  >
>  > Systems which want their skolem constants to be identifiable by other
>  > systems SHOULD use the .well-known URI prefix.
>  >
>  > - Steve
> 
> On 16/04/11 18:51, Steve Harris wrote:
>  > We will have to register the "genid" .well-known tag in any case,
>  > and tag: might have to be amended to say that .well-known is reserved 
>  > there too.

No need to amend tag:, just us a tag like
tag:w3.org,2001:bnode:[uuid-or-whatever]

> On 16/04/11 23:27, Pat Hayes wrote:
>  > It is not enough that *they* can so recognize it. It needs to be
>  > globally recognizable by any system that has access to the
>  > specifications. We need to specify how this can be done.
> 
> 
> I don't worry about dereferencability so prefer "genid:" -- currently 
> provide support for <_:...> to put bNodes in a different space to URIs 
> while reusing/abusing the syntax.

A key point about this proposal is that some people want
dereferenceability and some don't, so to move forward I think we need to
allow for both.   So folks who want to use http/https can, and folks who
want to use a non-referenceable scheme (genid:/tag:/bnode:/urn:bnode:)
also can.

   -- Sandro

>  Andy
> 
> FYI:
> 
> Registry for well-known:
> http://www.iana.org/assignments/well-known-uris/well-known-uris.xml
> 
> Registration list:
> http://www.ietf.org/mail-archive/web/wellknown-uri-review/current/maillist.html
> 
> 
>

Received on Wednesday, 20 April 2011 14:58:28 UTC