Re: Graphs / IRs / Blank Nodes - related strand

On Mar 4, 2011, at 3:27 PM, Nathan wrote:

> All, do note on Jonathan's diagram, only the bottom two boxes are critical, anybody can fight about the above two, but the fact is you can't see behind the interfaces or in to the abstract, so it doesn't matter about "states" and things like that from an HTTP/REST perspective, because it's hidden by the interface.

True, but we (RDF WG ) also have to connect all this to the RDF Concepts and Semantics, which are all about those abstractions. So it does matter for We Who Must Write the Specs. 

Pat

> 
> G-IR = IR, G-TEXT = Representation, all else, is hidden.
> 
> Pat Hayes wrote:
>> On Mar 4, 2011, at 3:05 PM, Jonathan Rees wrote:
>>> See attachment
>>> 
>>> Seems a bit unfortunate that they're taking the upper right route,
>>> while we're taking the lower left, but so it goes.
>>> 
>>> If they're claiming the G-SNAP is an IR...
>> For the record, in haste, my preference would be to say that the box is an IR, and the snap is an abstraction (in the sense that an RDF graph is a mathematical set and hence an abstraction) of the text, which is a representation of the instantaneous state of the box. Or, if you prefer, a representation of the box at a given moment in time.  Note that this allows 'boxes' which are indistinguishable from texts and which emit, when suitably poked, exact copies of themselves. So a text can be treated as a simple kind of box.
>> To express the relation of text to snap, one could use a phrase like 'an RDF text is required to parse into a legal RDF snapshot', ie the abstract graph is something like the RDF graph-parsing of the textual representation. The RDF model theory applies to snaps at present. The main point of my old ISWC talk about revising RDF can be phrased as the idea that blank nodes should always be understood as uniquely scoped to a text or box, i.e. two distinct boxes or texts cannot share a blank node. Pat
>>> well it's not clearly in
>>> agreement with TimBL.  I'm not sure I care as there is no way to tell
>>> the difference, as far as I can tell - the two boxes are isomorphic,
>>> and the diagram commutes.
>>> 
>>> Jonathan
>>> 
>>> On Fri, Mar 4, 2011 at 3:50 PM, Nathan <nathan@webr3.org> wrote:
>>>> FYI, a conversation Pat and I have been having on the RDF WG
>>>> 
>>>> http://lists.w3.org/Archives/Public/public-rdf-wg/2011Mar/0114.html
>>>> 
>>>> there's a lot more context, but that's about the top and bottom of it, named
>>>> graph (named g-box) that changes over time ("IR" then), blank node
>>>> identifiers scoped to the name where one exists, and where not then just to
>>>> the serialized rdf as per usual.
>>>> 
>>>> Nathan
>>>> 
>>> <g-stuff.pdf>
>> ------------------------------------------------------------
>> IHMC                                     (850)434 8903 or (650)494 3973   40 South Alcaniz St.           (850)202 4416   office
>> Pensacola                            (850)202 4440   fax
>> FL 32502                              (850)291 0667   mobile
>> phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
> 
> 
> 

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Friday, 4 March 2011 21:46:24 UTC