<kaz> scribenick: mleframc

<DanhLePhuoc> agenda: Agenda https://www.w3.org/WoT/IG/wiki/IG_Linked_Data_and_Semantic_Processing_WebConf

<inserted> scribenick: DanhLePhuoc

SSN-driven alignment proposal (presented by Maxime Lefrançois, Editor of SSN)

<inserted> scribenick: kaz

mlefranc: td pr 28
... wot td repository
... turtle document
... will go through it later
... second one is 348 on wot repo

<inserted> scribenick: DanhLePhuoc

two PRs for last assigned to :mlefranc

https://github.com/w3c/wot/pull/348

<inserted> scribenick: kaz

mlefranc: wot version of ontology used

<DanhLePhuoc> and https://github.com/w3c/wot-thing-description/pull/28

mlefranc: everything has a unique identity
... w3c.github.io/sdw/ssn/#SSNSystem
... sensor, actuator

<inserted> scribenick: DanhLePhuoc

mlefranc: is comparing side by side definition of Thing from SSN and TD

and the interaction patterns proposed by both

the Turle form of alignment of TD&SSN is proposed in https://github.com/w3c/wot-thing-description/pull/28

mlefranc: this alignment will import both SSN and TD

<kaz> https://w3c.github.io/sdw/ssn/#SSNProperty

<kaz> http://iot.linkeddata.es/def/wot#Thing

MariaPoveda: TD is developed by VICINITY project because it was needed to sake of time

then it was adopted by the group

it's some how reduced to minimal form

TD creator acknowledged there differences with SSN

kaz who is the main editor?

MariaPoveda: me

<kaz> (and some more contributors)

<inserted> scribenick: mlefranc

DanhLePhuoc: should alignments be part of normative or non-normative parts of the document (including examples) ?

<DanhLePhuoc> scribe mlefranc

MariaPoveda: I think the TD needs to be minimal, but we do not restrict alignments to other ontologies
... on the other hand the one to SSN should be standard,
... although maybe not directly in the ontology document

DanhLePhuoc: propose ideas and votes need to be part of a WG call isn't it ?

kaz: yes, but you can create an issue during this call, and explain it during the main call

Some strategies for ontology reuse on alignment with TD (presented by María Poveda)

MariaPoveda: <starting her presentation>
... three types of reusing ontologies:
... 1. import ontology (owl:import), 2. declare submodel of the ontology, 3. reuse the URIs of the ontology
... owl:imports is strong ontological commitment, you import all the axioms of the imported document, and this is transitive
... if the ontology server of the imported ontology fails, then everything breaks
... you can't delete/override any of the axioms in the imported ontology

2. declare submodel is copying definitions and axioms of another ontology,

scribe: a bit more difficult to implement, more robust to ontology server failures, less robust when changes in the imported ontology occur -> they are not propagated automatically
... for example in VICINITY our ontology copies the axioms and the definitions of the ssn-system module (see SOSA/SSN spec)
... 3. just reuse URIs of the reused ontology
... define your own concept, and link it to an external concept using axioms such as rdfs:subClassOf or owl:equivalentProperty, ...
... foe example, VICINITY extends the foaf:Person and defines: core:DigitalUser rdfs:subClassOf foaf:Person
... conclusions: combine 1, 2, and 3, stronger reuse when the external ontology is trustworthy (W3C > government/project > person)

DanhLePhuoc: would you make the TD ontology import any ontology at all ?
... ssn, schema.org, ...

MariaPoveda: maybe not directly in the TD ontology, maybe as mlefranc said: better in an external document
... I doubt that external ontologies (little or big) will be all implemented in TD implementations

DanhLePhuoc: I would suggest that we add example on how to use the TD ontology together with external ontologies
... we need some better synchronization between the WG and the IG
... with respect to TD ontology

Discuss on Ontology Alignment strategy for TD (focus on Linked Data perspective)

MichaelKoster: in iot.schema.org we are pretty well aligned with TD,
... in both cases we need at some point to use other ontologies to describe
... for example how a device acts on some feature of interest, etc.
... so let's keep new ontologies lightweight, and encourage the reuse of other existing formal ontologies
... maybe in a lightweight and reduced syntax such as JSON doc or so
... So: it's not so important to prevent terms conflicting in different namespaces
... the point is to see how these ontologies complement each other

DanhLePhuoc: iot.schema.org is impressive in how it combines different vocabularies and try to harmonize a terminology
... for domain ontologies conflicting on some terms
... we can't prevent people from developing new ontologies, the key is to help developers navigate through all the formal models, and choose the right way to model things
... let's create a taskforce with people from semweb, iot.schema.org, dev,

MariaPoveda: question is how to coordinate the addition of new semantics in the TD ontology
... what would be the freedom of this taskforce

DanhLePhuoc: maybe just propose new use cases and make them accepted first

kaz: basic mechanism of taskforces is that subgroup can make some decision in the taskforce call and bring it to the next call of the whole group
... for approval by the whole group

DanhLePhuoc: we haven't decided officially what we need to push to the main group, let's vote for the existence of the alignment and the fact it must be in an external document next week ?

<kaz> ACTION: maxime to clarify the proposal for the next IG call [recorded in http://www.w3.org/2017/09/01-wot-minutes.html#action01]

<trackbot> Created ACTION-116 - Clarify the proposal for the next ig call [on Maxime Lefrançois - due 2017-09-08].

[ adjourned ]