Re: Graphs and Being and Time

On Mar 1, 2011, at 6:14 PM, Nathan wrote:

> Pat, can you clarify for me, in your original mail, when you said "graph token" did you mean g-text (representation), or g-box (resource).

Um. At the time I hadn't made this distinction clearly. Sandro came along later and made it. I guess I ought to be honest and say that they are both 'tokens', and that at first I was thinking of something more like a text, and then thought that once we had tokens, they could also be things like boxes, which is a lot more use. 

> I'm asking because I'm trying to establish whether you're saying blank node identifiers would/could be scoped to the g-box (consistent over time) or per g-text (a property of the serialization, as currently).

I'd say both, but not to both a text and a box. That is, each text and maybe each box specifies a unique scope for any bnode. It can't be in two scopes at once. So when you take a snapshot of a box, you get a graph which is the 'state' of the box, but it has new bnodes in it. And each separate text, or even a copy of the text, contains its own unique bnodes. 

Pat

> 
> Best,
> 
> Nathan
> 
> Pat Hayes wrote:
>> On Feb 28, 2011, at 1:42 PM, David Wood wrote:
>>> On Feb 24, 2011, at 13:12, Pat Hayes wrote:
>>> 
>>>> It is much simpler: it is just wanting the WG to acknowledge that "an RDF graph" can either be a mathematical set, or it can be some kind of document or data structure or file than can be transmitted over a computer network. But it can't be both.
>>> What is the difference between an "RDF graph" and a RESTful "resource"?  
>> (Following definitions in existing specifications and discussions...) An RDF graph is a set, in the mathematical sense. This is a pure-mathematical abstraction, AKA a Platonic abstraction. A RESTful resource, as I understand it, is something that can be 'poked' with an HTTP (or other XXTP) request and when so poked will emit a "representation" of itself. I use scare quotes because exactly what this word means here has been the topic of much, um, discussion. But whatever it means, it is quite certain that emitting a representation of any kind is not something that a set could do, because sets can't DO anything. (And even if they could, they would do the same thing every time, whereas a REST resource is a thing with state, which can do different things every time, eg a clock.)
>>> What is the difference between an "RDF graph token" and a RESTful "representation"?
>> Well, a graph can have all kinds of tokens, to be strict, eg you can draw a picture of one on a piece of paper. But the kind I was talking about could be a RESTful resource, which was the main point. So whats the difference between a RESTful resource and a RESTful representation? Again, as I understand it, the key difference is that the first has a state - its like a g-box - whereas the second doesn't - its like a g-text (of the g-snap which is the abstraction defining the state of the g-box, if I have Sandro's terminology right.) Now, I admit that it would be perfectly correct to call a g-text a graph token as well. Sandro has made a further distinction that I wasn't trying to tackle, between the instantaneous 'graph state' of a g-box and the serialization of this graph that gets sent when you poke it with a transfer protocol prompt. Speaking purely from an abstract conceptual model point of view there isn't much to choose between these, but when we get into Web architecture it is a rather important distinction. Pat
>>> Regards,
>>> Dave
>>> 
>>> 
>>> 
>>> 
>> ------------------------------------------------------------
>> 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 Wednesday, 2 March 2011 05:32:53 UTC