Re: [SKOS] Re: Comments on SKOS Primer

Tom, thanks for your thoughtful comments on the emerging Primer--and
Antoine, thanks for responding in a more timely fashion than me :-)

I actually really like the basic/advanced distinction and agree it
would help people feel like they've got the essentials down pretty
quickly. Contrary to Antoine I think mapping concept schemes is kind
of an advanced topic. But I agree with Antoine that the Primer should
encourage SKOS users to link their concept schemes with other concept
schemes, before encouraging them to specialize SKOS itself. Perhaps
once the other reviews come in we can take a look at how the document
might look restructured in this way?

As for the graphs.... I consider myself a relative newcomer to semweb
technologies and can honestly say I'd rather look at some concise
n3/turtle over fancy graphs to get an understanding of an rdf graph.
But maybe my software development background biases me a bit in a way
that the Primer audience will not be?

I definitely think the graph + rdf/xml made sense for the Core Guide
[1]. I still find it difficult sometimes to get the sense of an rdf
graph with only the rdf/xml serialization in hand. Perhaps an overview
graph similar to what Alistair and Dan had in the Core Guide would be
a compromise?

For example the Primer could start off with a sample KOS snippet, then
present a visual representation of the equivalent SKOS graph, followed
by the rdf/xml and n3 serializations? This could potentially highlight
that SKOS isn't a particular type of picture or serialization but a
model that can be serialized in a variety of ways...and could set the
stage for the compact n3 or turtle that is used in the rest of the
document. Just a few thoughts at 3:30AM :-)

//Ed

[1] http://www.w3.org/TR/2005/WD-swbp-skos-core-guide-20050510/






On Jan 11, 2008 12:52 PM, Antoine Isaac <aisaac@few.vu.nl> wrote:
>
> Hi Tom,
>
> Honnestly I'm afraid I am also too much involved in it now to have a
> clear a valid opinion on these two issues (N3-only for examples is a
> problem, drawing a more precise line between "basic" and "advanced")
> right now.
> I would propose to discuss them a bit later, once we get more feedback
> from other reviewers or WG members (for instance, Justin seemed to have
> strong opinion on what was readable or not ;-).
> But clearly, let's keep them in mind.
>
> Antoine
>
> > On Thu, Jan 10, 2008 at 08:17:10PM +0100, Antoine Isaac wrote:
> >
> >>>   Using N3 presupposes that the audience for the Primer is
> >>>   more fluent in RDF than I had assumed.  My preference
> >>>   would be for the Primer to use visual graphs as in the
> >>>   2005 SKOS Core Guide [6].  I am assuming that graphs make
> >>>   it easier for readers who are new to RDF to see how things
> >>>   fit together.  In addition to the graphs, the 2005 guide
> >>>   also uses RDF/XML.  For the new Primer, my preference would
> >>>   be to keep using N3, though I wonder if all of these N3
> >>>   examples could be moved into an Appendix, shortening the
> >>>   body of the Primer (a good thing!).
> >>>
> >> That was more-or-less the initial plan to have graphs, and I have asked
> >> Alistair the sources for his graphs. However, due to lack of time I
> >> decided to drop them for the moment. Also, Ed remarked that this had the
> >> nice effect of reducing the length of the document (graphs usually takes
> >> more space), while keeping generally readable.
> >> I would propose to have a small test and do as you propose, but for a
> >> later version of the draft. What was true during the holidays remains
> >> true now :-(
> >>
> >
> > For now, we should perhaps just decide in principle whether
> > graphs should be used.  My concern is whether using N3 _only_
> > will be readable for the intended audience.
> >
> >
> >>>      In Advanced SKOS, conceptual resources can be _mapped_
> >>>      to conceptual resources in other schemes and _grouped_
> >>>      into labeled or ordered collections.  _Labels of concepts
> >>>      can be related_ to each other.
> >>>
> >> I will try to implement it in the version to review these days. But I
> >> don't guarantee that it will be ideal regarding the second part of your
> >> text. It does not fit the current structure of the Primer as well as it
> >> does for the reference :-( Unless we put "Advanced" SKOS in the title of
> >> both the "Networking" section and the "when KOS are not simple anymore"
> >> one. But I don't really like it: to me networking KOSs (re-using Concept
> >> Schemes, mapping, subject indexing) is less advanced and will be more
> >> common than things like grouping into collections and relationships
> >> between labels.
> >>
> >
> > That's a good point.  I saw the Essential/Advanced distinction
> > and liked it but have no strong opinion on where to draw the
> > line.  My point is more that if we do distinguish "advanced"
> > features, we should do so consistently and try to make it
> > part of the high-level story.
> >
> > Tom
> >
> >
>
>
>
>

Received on Sunday, 13 January 2008 08:36:27 UTC