W3C home > Mailing lists > Public > public-sdwig@w3.org > March 2019

Re: [sdw] New project proposal: OWL Space (#1095)

From: Rob Atkinson <rob@metalinkage.com.au>
Date: Mon, 11 Mar 2019 12:43:12 +1100
Message-ID: <CACfF9LwaVoqOzPfwhVcEhbCQQjN-Rt_izrpX43+8LLTmhqa6mA@mail.gmail.com>
To: Frans Knibbe via GitHub <sysbot+gh@w3.org>
Cc: public-sdwig@w3.org
From observing people try to tackle spatial concepts in ontologies (and
other structures) and attempting to describe organisation of data with
QB4ST I would suggest that a "kitchen sink" approach would be hard to make
work - it will be too complex for most and too simple or narrowly scoped to
meet many needs.

I would like to consider the "competency questions" for such an ontology
and specifically how it relates to others:

geoSPARQL defines a range of geometry data types and operations, but has
nothing to help describe what any such geometries might mean

many geometries are in fact observations of a feature property, and CRS is
technical detail of that result. Topology tends to be an easier case, but
only if you are not relying on spatial operations. competency questions
about how features relate topologically overlap _some_ of geosparql - but
we must remember that two features who are adjacent by definition may have
approximated geometries that do not have precise geometric relationship.

knowing what pieces we need to work together, but keeping each simple means
working out the competency questions for each part.  I would be looking for
symmetries here - as a straw man owl:time and owl:space may support similar
operations.  something that supports feature topology - topoSPARQL ? and
something that handles how to describe how a geometric property
semantically relates to a feature consistent with SOSA.

Rob












On Mon, 11 Mar 2019 at 07:00, Frans Knibbe via GitHub <sysbot+gh@w3.org>
wrote:

> Hello,
>
> It is great to see the idea being moved forward. And to a community group
> that is free for everyone to join too. I think that's very important.
> A concern could be that when something actually will be developed, it will
> be too heavily focused on geographic data. If that happens, the desire of
> aiding cross-domain interoperability could be at risk. And I really think
> that should be a core requirement, however small or modest the beginnings
> might be.  So it will be important to reach out to people that by nature do
> not flock to OGC community groups and might have never heard of the OGC in
> the first place. Luckily, the OGC is already in  a process of expanding its
> interests wider than basic geography, so that would not be breaking current
> practice, so it seems.
>
> Greetings,
> Frans
>
> --
> GitHub Notification of comment by Fransie
> Please view or discuss this issue at
> https://github.com/w3c/sdw/issues/1095#issuecomment-471338807 using your
> GitHub account
>
>
Received on Monday, 11 March 2019 01:44:24 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:17:53 UTC