W3C home > Mailing lists > Public > semantic-web@w3.org > June 2020

Re: RDF* and SPARQL* study

From: Stian Soiland-Reyes <soiland-reyes@manchester.ac.uk>
Date: Fri, 26 Jun 2020 09:57:27 +0100
Message-ID: <20200626095727.GD3351096@biggie>
To: "LJ.Garcia" <lj.garcia.co@gmail.com>
Cc: Antoine Zimmermann <antoine.zimmermann@emse.fr>, "Peter F. Patel-Schneider" <pfpschneider@gmail.com>, semantic-web@w3.org
On Thu, 25 Jun 2020 17:36:51 +0200, LJ.Garcia <lj.garcia.co@gmail.com> wrote:
> A possibility that could be considered if you see the responses to the
> questionnaire do not give you enough/uniform information, would be a
> webinar on the approach. That would allow for more discussion that could be
> useful for the research going on behind the questionnaire.

Yes, I would say the guidance document itself was top-notch, perfect for
initial teaching RDF/SPARQL/RDF*/SPARQL* concepts to students! Is it
available as a public web page?


I completed the survey and thought it was fun, I would like to see even
more of those "alternative but not always that good" modelling
approaches.


I am not so sure about the focus on property paths and inverses, but I
guess the hypothesis is that the usability/readability of these quickly
diminishes when the path gets complicated.


What I saw very little about was the RDF* modelling challenge of when
something should be reified or not - encapsulating statements hides
their facts behind a layer of obfuscation - that is good when the
statements are contradictory, but not so good when you just wanted to 
provide metadata.


For someone writing triples/queries it is also then becomes important to
know which facts are hidden within a reified triple, and what triples
are in the graph directly. How is this documentation captured when
ontologies and RDF Shapes have little RDF* support?


Reusing vocabularies can also be troublesome as there are not many
ontologies using Statement as subject or objects.

Other quad/graph-based approaches like Nanopublications [1] could be
used in comparison, or the n-ary relations with intermediaries like
schema.org Role [2] and PROV-O's use of the Qualified Patterns [3] - 
both of these are very much compatible with the property paths mentioned
in the survey, with the added benefit that you can identify and attach
more than one property to the statement.


As important as deciding HOW to use RDF* would also be the decision on
WHEN to use it. 


[1] http://nanopub.org/guidelines/working_draft
[2] http://blog.schema.org/2014/06/introducing-role.html
[3] https://www.w3.org/TR/prov-o/#description-qualified-terms


-- 
Stian Soiland-Reyes, The University of Manchester
https://www.esciencelab.org.uk/
https://orcid.org/0000-0001-9842-9718
  Please note that I may work flexibly – whilst it suits me to email now, 
  I do not expect a response or action outside of your own working hours.
Received on Friday, 26 June 2020 08:57:43 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:46:04 UTC