Re: Proposed agenda

With all due respect, regarding the memos you've directed us to, I
don't think it will be helpful at this point to talk about the
semantics of reference and binding in RDF.  I think we should try to
stick with normative RDF and OWL semantics and treat this
HTTP-semantics activity as an ordinary ontology development project.

When I make say things more vaguely and generally then they ought to
be said, I'm really trying to bait you all into fixing what I've said,
so that we have something to talk about and so that it's not just me
talking.  Maybe I do this because I'm not spending enough time on this
activity as I need to, but I really do hope my open-ended questions
are seen as drawing you out, and not an exercise in
gee-what-is-he-thinking.

I was therefore disappointed that in talking about the 'conundrum' you
didn't come up with your own classes - particular interpretations of
the vague categories I listed - for comparison with a putative IR
class. You did say we should "think about whether those entities have
the characteristics of an IR" - that's a great idea, but then you
didn't suggest any characteristics of an IR.

So I propose to take up your question: What are the characteristics of
an IR?  Or more broadly, what *might* be, what *has to be*, what
*cannot be* the characteristics of an IR?  For example:

. Does there exist an IR that has mass, or charge? - We have answered
  that pretty definitely no. So far so good.

. Is there at least one IR that has a dc:author? - I think we agree
  that there is.

. Does there exist an IR that has a number of pages, or a page size? -
  I don't know, what do you think?  (Especially Tim, what do you think?)

. How about spelling errors? Can an IR have spelling errors?

. How about a domain, or a range? I would say functions and relations
  have domains and ranges, but IRs don't.

. How about a license?  Certainly Creative Commons believes that certain
  things with 200-responding server behavior can be licensed.

I'm not assuming any particular definition of IR.  Just pick one and
let's talk about it.

So I would suggest the following homework:

Come up with fifteen interesting properties (or property/value pairs)
in three groups:

  - Five properties that some IR definitely has.

  - Five properties that definitely no IR has, but that some other
    thing (in each case) has.

  - Five properties that are mysterious to you.

Then we will collate, and see if there are any intersections between
our various sets.


Jonathan


On Wed, Jun 11, 2008 at 4:52 PM, Booth, David (HP Software - Boston) <
dbooth@hp.com> wrote:

> At our next meeting I'd like to propose that we discuss my answer
> http://lists.w3.org/Archives/Public/public-awwsw/2008Jun/0004.html
> to Jonathan's PDF file conundrum:
> http://lists.w3.org/Archives/Public/public-awwsw/2008May/0018.html
>
> To get there, we should first discuss:
>
> 1. http://dbooth.org/2007/splitting/ , particularly sections:
>           http://dbooth.org/2007/splitting/#akt and
>           http://dbooth.org/2007/splitting/#httpRange-14
> Alan may complain that the AKT example should have modeled AKT as a class
> instead of an instance, and that may be true, but for the purpose of this
> discussion I'd rather we ignore that issue, since the reasoning still holds
> even if the example is imperfect.
>
> 2. http://dbooth.org/2008/irsw/slides.ppt , particularly the two-step
> identity mapping described in slides 5-8.
>
>
>
> David Booth, Ph.D.
> HP Software
> +1 617 629 8881 office  |  dbooth@hp.com
> http://www.hp.com/go/software
>
> Opinions expressed herein are those of the author and do not represent the
> official views of HP unless explicitly stated otherwise.
>
>