W3C home > Mailing lists > Public > public-rdf-wg@w3.org > April 2012

Re: "layers" (was Re: the term "named graphs")

From: Dan Brickley <danbri@danbri.org>
Date: Sat, 28 Apr 2012 17:41:58 +0200
Message-ID: <CAFfrAFpxY-7SdPJpEHyS2rXtTz09xcFUaRdo2a3uFEeNCOyxQg@mail.gmail.com>
To: Sandro Hawke <sandro@w3.org>
Cc: RDF-WG Group <public-rdf-wg@w3.org>

Glad 'layer' appeals to you. I think I picked it up from the mcf spec
originally, http://www.guha.com/mcf/wp.html where it has some more
specific meaning (layers are ordered, and triples are true-or-false
w.r.t. a layer). Or the old Mozilla APIs which also had this. But
Pat's talk of 'surfaces' a few while back was in a similar direction;
triples/claims are written on a surface. Not sure how far we can push
this but the metaphor seems potentially quite developer friendly...

I tried drawing this a few ways, e.g.
http://www.flickr.com/photos/danbri/3472944745/ where each layer there
has a hint as to provenance. Or felt-tip-pen-version,
http://farm4.static.flickr.com/3613/3384528143_8304792836_b.jpg ...

And yes, you're right that it's perhaps a bit weird that URIs are 'on'
multiple layers. Maybe we say (per Pat below) they're "written" on
multiple layers?

files are here,
with triples and sparql examples nearby,

e.g. (some goofy uris),

"#According to the layerlist, who made the pages that tell us a
schoolHomepage for Alice?
PREFIX : <http://xmlns.com/foaf/0.1/>
SELECT ?g ?who WHERE {
  GRAPH  <http://localhost/notube/layerlist.rdf> {  ?who :made ?g . }
  GRAPH  ?g {    <http://localhost/notube/layer1.rdf#alice>
:schoolHomepage  <http://lookingglass.example.org/> .  }

For surfaces, see http://www.slideshare.net/PatHayes/rdf-redux ... transcribing:

"""RDF graphs are drawn on surfaces. Blank nodes are marks on the
surface. intuitively, think of a surface as a piece of paper, or a
screen, or a document. .... Surfaces provide the missing type/token
distinction. Putting the same graph onto a new surface is like making
a copy. But copying a graph onto a new surface always gets you new
blank nodes, because a mark can only be on one surface""".

This might be out of date w.r.t. Pat's current thinking but I think
the metaphors are heading in very similar directions.

To pick on the bnode issue, we could imagine operations like copying
from one layer/surface to another, which merits a new bnode. Or maybe
also stitching two layers/surfaces together in a way that preserved
bnodes. (Does modern sparql allow the 'same' bnode in two named
graphs? sorry I forget!)

The only terminology clash I see is with 'layer' from 'layercake'; but
that diagram is fairly internal to the semweb community, and we're
needing terminology that 100s of 1000s of web developers can embrace.

> One cool thing is how much it raises the question, "layer in what?"

Yes, I like that. Not sure how precise we can be; do we consider Web
pages layers, or just somehow things that can be intimately related to

Received on Saturday, 28 April 2012 15:42:49 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:02:04 UTC