On 10/22/2013 08:26 PM, Pat Hayes wrote:
David, a quick response to correct a misapprehension. See inline
below.
On Oct 22, 2013, at 9:05 AM, David Booth wrote:
Hi Peter,
On 10/21/2013 06:48 PM, Peter F. Patel-Schneider wrote:
Is this request supposed to be for me, or for the sender of the
response? I initially sent back a private response on this, but
in the interests of time, I will answer with my personal
feelings.
Yes, but apparently you missed my followup, as I didn't receive a
response to that. My followup was: [[ For concepts that are *used*
then I would agree, but that concept is *not* used in the RDF
Concepts spec. The RDF Semantics spec uses other far more
important concepts too, such as "denotes", but surely you would not
advocate moving those definitions to the RDF Concepts document? ]]
The introduction of generalized RDF is in Concepts because
Concepts is where RDF concepts are to be introduced.
Generalized RDF was called out as a worthy RDF concept because
JSON-LD needed something to point to for its generalization of
RDF.
And my followup said: [[ That's an interesting catch-22, because
the JSON-LD *justification* for using the notion of generalized RDF
was that it is defined in the RDF specs, so we seem to have a
circular justification going on here.
No, the reason for JSON-LD using a generalization of RDF syntax is
quite external to the RDF specs, and has its roots in the JSON
community itself. So this is not a catch-22, as you put it, but a
small gesture of conciliation between RDF and JSON-LD, which are like
two musicians playing the same tune but each insisting that their
version was written by a different composer.
I understand that the *motivation* for using generalized RDF had nothing
to do with the existence of the definition in the RDF specs. But i
distinctly remember the existence of the definition in the RDF specs as
being used as a justification for going ahead with the
blank-nodes-as-predicates feature in JSON-LD, rather than finding
another solution in spite of the motivation. I don't know how easily i
could find the messages or phone log to prove that, but i was quite
heavily involved in those discussions at that time and i distinctly
remember that point standing out in my mind at the time.
David
Pat Hayes
In what sense do you view the Concepts document as being a better
reference than the Semantics document? Are you suggesting that the
definition *should* have more prominence than it would get in the
Semantics doc? The problem with giving it more prominence is that
people start to misconstrue it as being a W3C standard on par with
standard RDF. But generalized RDF has not gone through at all the
same level of rigor as standardized RDF -- no test cases, no
interoperable implementations, etc. -- and was not intended by the
W3C to be promoted as a W3C standard. The fact that JSON-LD
references that definition is a bug, not a feature, IMO. ]]
Bottom line: I'm not at all convinced by the rationale that I've
heard so far, that the Concepts document is a better place for this
definition than the Semantics document. Is there more rationale
that I've missed? Or do you disagree with my points above? If so,
what and why?
David
peter
On 10/16/2013 10:10 AM, David Booth wrote:
Hi Peter,
The wording of this definition looks good to me, but why are
you opposed to moving it to the RDF Semantics document?
AFAICT, the term is not used in the RDF Concepts document, but
it *is* used in the RDF Semantcs document. Also, moving it to
RDF Semantics would give it less visibility, which (to my mind)
would be appropriate given that standard RDF is what the W3C is
intending to promote, rather than generalized RDF.
David
-------- Original Message -------- Subject: Re: RDF Concepts -
Definition of "Generalized RDF" Resent-Date: Wed, 16 Oct 2013
13:11:52 +0000 Resent-From: public-rdf-comme...@w3.org Date:
Wed, 16 Oct 2013 09:11:18 -0400 From: David Wood
<da...@3roundstones.com> To: David Booth <da...@dbooth.org> CC:
RDF Comments <public-rdf-comme...@w3.org>
Hi David,
This is an official response from the RDF Working Group
regarding your comment at [1] on the definition of "Generalized
RDF". Your comment is being tracked at our ISSUE-147 [2].
The WG discussed your concerns at our 2 Oct telecon [3] and via
email [4]. Those discussions resulted in a decision to leave
the definition of "generalized RDF" in RDF 1.1 Concepts, but to
change the definition to the following: [[ Generalized RDF
triples, graphs, and datasets differ from normative RDF
triples, graphs, and datasets only by allowing IRIs, blank
nodes and literals to appear anywhere as subject, predicate,
object or graph name. ]]
My action to make the editorial changes was tracked at [5].
The updated section 7 is available in the current editors'
draft [6].
Please advise the working group whether this change is
acceptable to you by responding to this message. Thank you for
your participation.
Regards, Dave -- http://about.me/david_wood
[1]
http://lists.w3.org/Archives/Public/public-rdf-comments/2013Oct/0006.html
[2] ISSUE-147: https://www.w3.org/2011/rdf-wg/track/issues/147
[3] https://www.w3.org/2013/meeting/rdf-wg/2013-10-09#line0228
[4]
http://lists.w3.org/Archives/Public/public-rdf-wg/2013Oct/0030.html
[5] ACTION-309: https://www.w3.org/2011/rdf-wg/track/actions/309
[6]
https://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html#section-generalized-rdf
------------------------------------------------------------ IHMC
(850)434 8903 home 40 South Alcaniz St. (850)202 4416
office Pensacola (850)202 4440 fax FL
32502 (850)291 0667 mobile
(preferred) pha...@ihmc.us http://www.ihmc.us/users/phayes
