Re: Ordering concepts in a Tree display

Hi Enzo,

While I certainly agree that ordering OFTEN carries semantics, there are
cases where the order in which the resources are displayed to the user
simply does not. The immediate case for us is replication of the order that
the resources we're displaying appeared in the original unstructured
document (we're expressing structured resources contained in unstructured
textual documents). Sometimes the document order provides information
necessary to the correct understanding of the resources and the relationship
between the resources and sometimes it does not. When it does, we capture
the order. Another case for us is language-specific alphabetizing of all or
portions of a graph, which makes the display more readable for humans, but
conveys no additional meaning to machines consuming the data.

That said, I think the original post regarding 'natural' ordering of
concepts is a case, like yours seems to be, where it's useful to formally
express the meaning of 'natural order' in a way that explicitly states the
knowledge behind the data creator's understanding of the ordered
relationship between concepts (historical period, year invented, etc.) in a
way that can also be used to express that order in a display at a human
endpoint. SKOS was quite intentionally designed to be extended in this way.

I'd also like to observe that discussions like this often take an
understandably RDF-only, triple-store-centric view of data modeling and
management that can make relatively simple problems fiendishly complex. RDF
neatly solves many problems but often too, the data modeling requirements
differ for creation, dissemination, aggregation in an open world, indexing,
and display of metadata, and this can present distinct and often unnecessary
challenges when RDF is the only tool available.

I quite like your paper, by the way, particularly your useful and detailed
description of your methodology in defining post-coordination facets.

Cheers,
Jon


On Fri, Feb 4, 2011 at 6:21 AM, Maltese Vincenzo (Enzo) <
maltese@disi.unitn.it> wrote:

> Dear Jon,
> I'm new to this community, but let me express my viewpoint about this
> issue.
>
> In fact, while we agree (my colleague Bisu in CC and I) that ordering is
> purpose oriented, it DOES carry semantics.
> As it is explained in the attached paper to appear at GEOS 2011, it gives
> implicit semantic relations between coordinate terms. This can be of great
> value for usability issues when a user browses a classification scheme.
>
> Bests,
> Enzo
>
>
> -----Original Message-----
> From: public-esw-thes-request@w3.org [mailto:
> public-esw-thes-request@w3.org]
> On Behalf Of Jon Phipps
> Sent: Monday, January 31, 2011 9:17 PM
> To: Alistair Miles
> Cc: Tom Morris; Jakob Voss; SKOS
> Subject: Re: Ordering concepts in a Tree display
>
>
> Alistair Miles wrote:
> > (2) represent your systematic display using some other data format
> > (some sort of XML would be ideal, as you get hierarchy and ordering
> > easily), in which case you'd have to figure out how to manage your
> > systematic display data in addition to your basic broader/narrower
> > graph and make sure the two weren't inconsistent,
>
> Hi Alistair,
>
> This is the approach we're taking in our upcoming refresh of the Registry,
> since we believe that in general ordering for display is a 'local' system
> issue rather than an expression of conceptual semantics and is liable to be
> highly variable. Not to mention that ordering for human browsing is usually
> completely unrelated to semantics.
>
> We're using JSON to express a 'manifest' that can be easily displayed as an
> ordered hierarchy, the ordering of branches and leaves performed and stored
> independent of the maintenance of the RDF.
>
> Nice to have you around,
>
> Jon Phipps
> http://metadataregistry.org
>
>


-- 
Jon

I check email just a couple of times daily; to reach me sooner, click here:
http://awayfind.com/jonphipps