Re: Mappings between SSN & WoT TD

Dear all,

Following Lionel Medini's request to provide an initial alignment between
the SOSA/SSN and the WoT ontology (see mail below),

Considering the TD JSON-LD context are at
 - http://w3c.github.io/wot/w3c-wot-td-context.jsonld
 - http://w3c.github.io/wot/w3c-wot-common-context.jsonld

and considering that these contexts refer to the ontologies at
  "td": "http://iot.linkeddata.es/def/wot#", and
"saref": "http://uri.etsi.org/m2m/saref#", (shouldn't it be
https://w3id.org/saref# ?)

I would rather open the discussion to Raul, Maria and Victor. My initial
guess for an alignment between the SOSA/SSN ontology and the VICINITY wot
ontology would be:

wot:Thing rdfs:subClassOf ssn:System .
wot:providesInteractionPattern rdfs:subPropertyOf ssn:implements .
wot:InteractionPattern rdfs;subClassOf sosa:Procedure .
wot:hasInputData rdfs:subPropertyOf ssn:hasInput
wot:hasOutputData rdfs:subPropertyOf ssn:hasOutput


@authors of the wot ontology, some preliminary comments:
- I just noticed what may be typos in the definition of wot:MediaType:
estructured --> structured, os --> of , Definicion --> Definition
- typo in the label of wot:isReadableThrough
- I would suggest to rename wot:Property to wot:PropertyInteractionPattern
to avoid confusion with ssn:Property
- same for wot:Event and wot:Action ?
- do you really want to impose the use of OM for units of measures ? or
could we use QUDT instead ?
- instead of wot:DataSchema, couldn't we use the class
rdfp:GraphPresentation from the RDFP ( https://w3id.org/rdfp/ ) ontology to
generalize a bit to any RDF Graph that has some validation rules / lifting
rules / lowering rules ? That could help to cover cases where input data or
output data do not solely consist in a quantity value  (ex. some text,
concepts, or force and torque values)
- could we have wot: properties that map to the CoRE resource directory rt
(resource type) and if (interface) Web Link target attributes ? see RFC6690
- instead of a property with a boolean range, I've heard it's good practice
to use classes instead --> disjoint classes RequiredProperty and
OptionalProperty ?

Some other suggestions come to my mind, but that should be a good starting
point to develop discuss further the development of that nice wot ontology
and it's alignment to SSN.

@Lionel, some more comments inline

Best,
Maxime Lefrançois

 Le ven. 7 juil. 2017 à 16:41, MEDINI LIONEL <lionel.medini@univ-lyon1.fr>
a écrit :

> Hi Maxime,
>
>
>
> As I understand, Danh won’t have time to answer this email, so I rely on
> you.
>
>
>
> Currently, I have mapped sosa:Platform as subclass of wot:Thing.
>

Because I would like to align ssn:implements with
wot:providesInteractionPattern, then I suggested to align ssn:System to
wot:Thing instead.


> In order to show requests that are able to do more complex things than
> retrieving instances of WoT Thing, I will need more mappings between the 2
> ontologies. For instance, it would be good if we could retrieve all things
> in a given area (deployment ?),
>

Deployment should not be considered here, you can add lat/long coordinates
to anything provided that it is physical --> sosa:Platform, sosa:Sensor,
sosa:Actuator, sosa:Sampling, ssn:System... (not sure about ssn:Deployment
because it's aligned to dul:Event....)


> or a list of available temperature sensors by querying TD classes and
> properties.
>

>
> Do you have / could you provide me with a turtle file stating such
> mappings?
>
>
>
> Thanks in advance for your help,
>
>
>
> Lionel.
>
>
>
> --------------------------------------
>
> Lionel Médini - associate professor
>
> LIRIS Lab / University of Lyon
>
> Phone: +33 4 72 43 16 36 <04%2072%2043%2016%2036>
>
> Fax: +33 4 72 43 15 36 <04%2072%2043%2015%2036>
>
> mailto:lionel.medini@liris.cnrs.fr <lionel.medini@liris.cnrs.fr>
>
> https://liris.cnrs.fr/lionel.medini/
>
>
>