Fwd: some feedback about the graph concept terminology

Hi All, 

I sat with Luc Moreau one of the chairs of the Provenance WG a couple of days ago and he has sent us some feedback on the Graph Concept Terminology which I am forwarding below. 

Mischa

Begin forwarded message:

> From: Luc Moreau <L.Moreau@ecs.soton.ac.uk>
> Date: 13 April 2011 09:39:15 CEST
> To: mischa.tuffield@garlik.com, Sandro Hawke <sandro@w3.org>, Paul Groth <pgroth@gmail.com>
> Subject: some feedback about the graph concept terminology
> 
> 
> Hi,
> 
> Mischa pointed me to the Graph Concept Terminology. I understand this
> will be discussed at your F2F meeting in Amsterdam.
> 
> I find the distinctions between the notions of g-box/g-snap/g-text
> clear.  They map naturally to the notion of stateful web resource,
> content negotiation, and state serialization.
> 
> 
> With a provenance hat on, there are a few things that we need to
> consider.
> 
> 
> 1. The document says:
>   "Like g-boxes, g-snaps can overlap, sharing nodes and arcs. Unlike
>    g-boxes, it makes no sense to talk about g-snaps changing: they are
>    defined to be exactly the collection of their elements."
> 
>   For those interested in provenance, we may have two identical sets
>   (in the mathematical sense, i.e., with the same elements),
>   but it does not mean that these sets were produced in the same way.
> 
>   Hence, I would like to be able to name each g-snap, so that I can say
>   that the provenance of g-snap1 is prov1 and the provenance of g-snap2
>   is prov2, even though both g-snap1 and g-snap2 have the same triples.
> 
>   Coming back to the Web architecture, it would be easy for the
>   g-box, to mint new URIs for each g-snap being created.
> 
>   This may be captured by "Possible requirement 2", but I wasn't sure.
> 
> 
> 2. While this proposal addresses a number of use cases, I am not sure it covers
>   all the needs to "name graphs".
> 
>   In the Open Provenance Model, there is a notion of account, i.e. a subset
>   of a provenance graph, containing a set of provenance assertions which are
>   "logically connected" (for instance, assertions made by a same observer,
>   assertions coming from a query and merged in another graph, or assertions
>   describing provenance at some level of abstraction).
>   (Note that there is some desire to have a notion of account in a
>   standardized Provenance Interchanged language, but its actual shape
>   needs to be finalized.)
> 
>   Hence, such a provenance account does not necessarily correspond to
>   a g-snap. Instead, it's more like a subset of a g-snap.  Is there a
>   mechanism to name a subset of triples inside a g-snap?
> 
> I hope this is helpful,
> Feel free to forward my email to the rdf mailing list if appropriate,
> Cheers,
> Luc
> 
> 
> 
> -- 
> Professor Luc Moreau
> Electronics and Computer Science   tel:   +44 23 8059 4487
> University of Southampton          fax:   +44 23 8059 2865
> Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
> United Kingdom                     http://www.ecs.soton.ac.uk/~lavm
> 

___________________________________
Mischa Tuffield PhD
Email: mischa.tuffield@garlik.com
Homepage - http://mmt.me.uk/
Garlik Limited, 1-3 Halford Road, Richmond, TW10 6AW
+44(0)208 439 8200  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD