the choice of the term "layer"

On Wed, 2012-05-02 at 17:44 +0200, Dan Brickley wrote:
> On 2 May 2012 16:47, Guus Schreiber <> wrote:
> > On 02-05-2012 15:42, Sandro Hawke wrote:
> >> I took the liberty of moving forward with drafting a possible spec, so
> >> we have something to look at.   Several sections are empty, but I'm
> >> pretty happy with what's there.  The example is that same as on the
> >> "Layers" page on the wiki.
> >
> > Terminology issue (I would by no means want to disturb any arising
> > consensus).
> >
> > I don't think the term "layer" will do the required trick. I cannot but
> > associate it with vertical relations. The term we choose should have both
> > vertical and horizontal connotations. I'd prefer "box": boxes can be put
> > next to each other or on top of each other.
> >
> > Feel free to ignore for the moment.
> RDF data can also be stitched together into a single flat thing, like a quilt.
> The main reason I like 'layer' (and surface) is that data integration
> is RDF's defining and most under-sold feature, and this metaphor does
> highlight that feature...

I agree.    Unfortunately, when I looked at the title "RDF Layers" with
fresh eyes, I immediately thought of the SemWeb layer cake.  :-(   So, I
put in "Data", sometimes calling them "Data Layers" to help with that.

In the draft I currently have it flagged as an open issue, with some
other possibilities listed.  The one I woke up with this morning was
"spaces".   Conceptually it's close enough to tuple spaces that I think
the similarity in names would probably be okay.  (I just learned what
Linda was named after...!  Who knew...?)

(Downside to "space" is that is start with "s", so quads are (S,P,O,S)
which is annoying.)

   -- Sandro

Received on Wednesday, 2 May 2012 16:15:32 UTC