[SKOS] Standalone definitions in natural language; previous links in headers


On Tue, Mar 18, 2008 at 02:56:54PM +0000, Alistair Miles wrote:
> >>I imagine the directionality is somewhere stated, but in searching 
> >>throught the SKOS Reference for several minutes I was unable to find it 
> >>in the obvious places, such as the section on skos:broader:
> >>http://www.w3.org/TR/2008/WD-skos-reference-20080125/#broader
> >>or in Section 7 on its semantics:
> >>http://www.w3.org/TR/2008/WD-skos-reference-20080125/#L2055
> >>or even in the examples:
> >>http://www.w3.org/TR/2008/WD-skos-reference-20080125/#L2157
> >>
> >>I think a statement needs to be added somewhere, even if it is merely in 
> >>the example, saying something like:
> >>[[
> >>        :a skos:broader :b .
> >>
> >>indicates that :b is a broader concept than :a.
> >>]]

> >I assume the entries in Section 13 of SKOS Reference [1]
> >will need to show Definitions, as in the 2005 spec [2] (and,
> >as in 200, perhaps a clickable index of Classes and Properties
> >in the upper right corner).
> I had no intention to add definitions to these tables. I think the 
> sections of the main body of the document itself serve perfectly well as 
> definitions, there is no need to create standalone definitions for each 
> class or property.
> I'm reluctant to work on separate definitions because agreeing on short 
> standalone definitions can be very hard, and could take time we don't 
> necessarily have to spend.

I see that the RDF vocabularies [8,9] use natural-language
definitions but the OWL vocabulary [10] does not.

In the case of SKOS, not using natural-language definitions
would mean not creating definitions for new properties, such
as skos:seeLabelRelation, but it would also mean throwing
out definitions used in the 2005 specification [2].

Definitions would provide a ready-reference answer to David's
original question (see above), who wanted to know what is
broader than what with skos:broader.

I know from experience with Dublin Core how much work it can
be to agree on short standalone definitions but wonder if it
is a good idea simply to dispense with them, and whether it
sets the right precedent for other such vocabularies, especially
given the broad audience that SKOS is intended to reach.

> >David's observation makes me notice that the current SKOS
> >Reference [3] lacks a Previous version: link.  As the Latest
> >Version: link of the older SWBP draft [4] now redirects to
> >[3], need there not be an unbroken chain of Previous version:
> >links from [5] to [6]?  (Or is it the link from [4] to [3]
> >that is in error?)
> There is a note about superseding the older draft (with a link) in the 
> SOTD section.

Other Working Drafts, such as RDFa Syntax [7], have Previous version:
links in the headers.  Previous version: links are used in the example
page [12] I generated using the pubrules filter [11] - maybe Ralph
can say if these links are actually required?


> >[1] http://www.w3.org/TR/2008/WD-skos-reference-20080125/#broader
> >[2] http://www.w3.org/TR/2005/WD-swbp-skos-core-spec-20051102/#broader
> >[3] http://www.w3.org/TR/skos-reference/
> >[4] http://www.w3.org/TR/swbp-skos-core-spec
> >[5] http://www.w3.org/TR/2008/WD-skos-reference-20080125/
> >[6] http://www.w3.org/TR/2005/WD-swbp-skos-core-spec-20051102/

[7] http://www.w3.org/TR/2008/WD-rdfa-syntax-20080221/
[8] http://www.w3.org/1999/02/22-rdf-syntax-ns#
[9] http://www.w3.org/2000/01/rdf-schema#
[10] http://www.w3.org/2002/07/owl#
[11] http://www.w3.org/2005/07/pubrules
[12] http://www.w3.org/2005/07/pubrules?year=2008&uimode=filter&filter=Filter+pubrules&filterValues=form&docstatus=ord-wd-tr&patpol=w3c&rectrack=yes&normative=yes&prevrec=none#docreqs

Tom Baker - tbaker@tbaker.de - baker@sub.uni-goettingen.de

Received on Tuesday, 25 March 2008 14:15:04 UTC