action: choices for anonymous resources

In the July 13 minutes, I took an action to:  "Clarify the choices for
anonymous resources with some of their ramifications".  This is
essentially a reprise (with some amplification) of the points I made in
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0124.html 
(quoted in Graham's response
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0131.html)

In my earlier message, I said:

"I think the main (and fundamental) argument for anonymous resources in 
the model/abstract syntax is to support, in the model/abstract syntax, 
the capability described in the current M&S for talking about resources 
that don't have URIs (that you know of), e.g.(P41):

"The individual whose name is Ora Lassila, email <lassila@w3.org>, is 
the creator of http://www.w3.org/Home/Lassila."

the graph for which shows a resource *without a URI*.

That is, the current M&S says we can express this in RDF.  I interpret 
this as meaning that the model/abstract syntax must be capable of 
expressing this, not just the XML serialization.  It seems to me, then, 
that the choice is whether to (a) explicitly support anonymous resources
as something other than resources with generated URIs;  (b) clarify this 
part of the spec by saying we really have to generate a URI in this case 
(and have the graph show a *generated* URI, not *no* URI);  (c) delete 
this capability from the spec."

On going back and reading the issues list (and some other things), I now 
realize there's a bit more to this than I originally thought (for one 
thing, both the current RDF container model and reification seem to 
involve anonymous resources too), but rather than dealing with those 
cans of worms in this message, I'll stick to trying to clarify the 
choices I described earlier, and talk about those other issues later. 
However, one bottom line here is that I agree with everyone who's said 
that we've got to clarify a lot of issues surrounding URIs as a matter 
of very high priority.

Now to the immediate issue:  We need to be clear about how we're 
interpreting "anonymous resources".  The graph pictures show them as 
having *no URIs*.  However, parsers produce generated identifiers for 
these things, and our triple model seems to require some such 
interpretation, that is, that these apparently "anonymous" resources 
have some identifier produceed for them that can be plugged into the 
appropriate places in the triples).  So my original choices for what was 
happening were (in the same order):

a.  these resources are "anonymous" in the sense that they don't have 
URIs.  However, they have to have *identifiers* of some kind in order to 
reflect them in triples.  In this case, the M&S should say that another 
kind of identifier is internally generated to identify "anonymous" 
resources, and say what the characteristics of those identifiers are 
(like, they're internal, temporary, not meant for use outside of 
something like a "session", ...) and how they work compared to ordinary 
URIs.

b.  these resources are "anonymous" only in the sense that the writer of 
the RDF doesn't have to explicitly supply a URI.  However, a genuine URI 
will be generated for the resource, and once generated this URI will 
behave like any other URI.  In this case, the M&S should say explicitly 
that URIs are generated for "anonymous" resources, the graph diagrams 
should show *generated* URIs, not *no* URIs, for these resources, and 
the M&S should say how these generated URIs work (e.g., when different 
parsers operate on the same XML serialization).

c. we don't want to support generated identifiers, either some special 
kind of non-URI identifier, or generated URIs.  Then we should remove 
"anonymous" resources from the M&S (e.g., require explicit specification 
of URIs in these cases, hand-wave about how URIs get assigned in a 
manner outside the scope of the M&S, or some alternative TBD).

Note the distinction I'm making in choices (a) and (b).  I understand 
that parsers generate *identifiers* for anonymous resources.  What I 
want to be clear about (if only in my own mind) is *what kind* of 
identifiers these are expected to be.  So, for one thing, in future 
discussion it would be helpful if we wouldn't use generic terms like 
"genids" or "generated IDs" for these things if what we mean is 
"generated *URIs*" (of course, clearing up URI semantics would help 
clarify the requirements of "generating URIs" in such circumstances).

--Frank

-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875

Received on Tuesday, 17 July 2001 17:52:42 UTC