Re: New proposal for SOSA

Hi Maxime,

Thanks for proposing a model for the core ontology.

It would be useful if you can relate back your changes/additions to the ontology to issues that have already been raised on the tracker in the current version of SOSA. If there are further changes that you would like to see, please add them to the tracker.

I don’t think anyone has the bandwidth to compare your proposal to the one that has evolved over the last seven months on Github: https://github.com/w3c/sdw/blob/gh-pages/ssn/rdf/sosa.ttl


Thanks,
Armin

From: Maxime Lefrançois <maxime.lefrancois@emse.fr>
Date: Tuesday, 7 February 2017 at 1:09 pm
To: Armin Haller <armin.haller@anu.edu.au>, "janowicz@ucsb.edu" <janowicz@ucsb.edu>, "Simon.Cox@csiro.au" <Simon.Cox@csiro.au>, Kerry Taylor <kerry.taylor@anu.edu.au>, "public-sdw-wg@w3.org" <public-sdw-wg@w3.org>
Subject: New proposal for SOSA

Dear all,

Please find attached a proposal for SOSA, and a figure that illustrates it.

1)

I have taken the liberty to add some RDFS axioms in SOSA, knowing that such axioms will never prevent SOSA to become one day part of schema.org<http://schema.org>. As a matter of fact, goodrelations, the well known ontology that has been integrated into schema.org<http://schema.org> is highly axiomatized:
 - it uses disjoint classes axioms
 - it uses domains and ranges that are union of classes.

The second point makes it pretty well aligned with schema.org<http://schema.org> domainIncludes and rangeIncludes actually, while using only the RDFS and OWL vocabulary.

See http://purl.org/goodrelations/v1.owl



2)

I also propose to remove the sosa:hasValue.

schema.org<http://schema.org> already has its way of assigning values to elements (see http://schema.org/value). Plus, other ways exist to give a value and a unit of measure to an observation result (QUDT, OM,..)


3)

I have renamed some properties,
 - either to be more aligned with SSN
 - or so that there is a clear naming convention


There may be other noticeable differences that I did not document yet.


Shall I issue a pull request before we discuss these different points in separate threads ?



I'm working on the rest of my proposal:
 - SSNX: http://www.w3.org/ns/ssn/ssnx

 - SSN: http://www.w3.org/ns/ssn/

 - various alignment documents proposed by Simon
 - some specifications for the server to expose documents and terms in accordance with best practices.

The old namespace http://purl.oclc.org/NET/ssnx/ssn#<http://purl.oclc.org/NET/ssnx/ssn> would redirect to http://www.w3.org/ns/ssn/ssnx


 - Ontology SSNX deprecates the old terms and aligns them with either a sosa term, or a ssn term
 - Ontology SSN imports SOSA, adds axioms, and other SSN terms.


To me, the only tricky parts in the integration of SOSA/SSN so far are:
 - property oldssn:isProducedBy
 - SensorOutput and Observation are merged into a single class.

Best,
Maxime