Spatial Data on the Web Working Group Teleconference
30 May 2017
<Kjanowic> +1
+1
<ahaller2> +1
<mlefranc> +1
<DanhLePhuoc> +1
Patent call
Markus Stocker feedback resolved and merged.
Comments provided from Nick Car (GA) -
... re feature of interest
mlefranc: "platform" used slightly differnet in WoT
… WoT may need additional class
Kjanowic: sosa:Platform supports composition of sensors - e.g. smartphone
… or if need to talk about geometry
(WoT needs 'platform' for software platform)
is platform just *hardware* platform
mlefranc: also asking issue about sub-system
<Kjanowic> Sorry, Id did not mean that. We include virtual as well. My point was that it is not an API platform
in IT world platform=API
in SOSA we mean host for system, sensor, but not API platform. Not restricted to hardware
ahaller2: don't need to change definition/annotation
<Kjanowic> Wintel is a platform but not in our sense
DanhLePhuoc: must delegate to WoT group - they are still struggling around what is central point of modelling
… we don't need to change.
… wot still a moving target
… haven't agreed thing-description in JSON, JSON-LD yet ... difficult to align until they have stablised
<DanhLePhuoc> +1
mlefranc: best explained through examples
… subsystem vs ishostedby
… also isSampleOf and hasSubSystem
<ahaller2> create ISSUE What is the difference between subsystem vs ishostedby
<mlefranc> ISSUE What is the difference between subsystem vs ishostedby
ISSUE: What is the difference between subsystem vs ishostedby
<trackbot> Created ISSUE-202 - What is the difference between subsystem vs ishostedby. Please complete additional details at <http://www.w3.org/2015/spatial/track/issues/202/edit>.
isSampleOf is to get a sub-feature which is intended to be 'representative of' a bigger thing in the context of observations
Action: SimonCox to add text to explain role of sampling in SSN examples room/ice sheet
<trackbot> Created ACTION-364 - Add text to explain role of sampling in ssn examples room/ice sheet [on Simon Cox - due 2017-06-06].
<Kjanowic> This is a dangerous example btw "<Room145/east> a sosa:Sample , sosa:FeatureOfInterest , sosa:Platform ;"
Kjanowic: danger of mixing sample/foi/platform in examples
… don't conflate too many things in examples as it is a rare case
… need to be careful about conflating sampling locations
<Kjanowic> Sorry for jumping ahead
<Kjanowic> but it relates to sample vs system and foi
<Kjanowic> yes, absolutely
<Kjanowic> but the example is, correct?
<Kjanowic> <Room145/east> a sosa:Sample , sosa:FeatureOfInterest , sosa:Platform ; rdfs:label "East wall of room #145."@en ; rdfs:comment "This wall hosts PCB Board 1 with DHT22 temperature and humidity sensor #4578."@en ; sosa:hosts <PCBBoard1> .
<Kjanowic> Thanks Maxime for providing these examples
<Kjanowic> yes to very simple examples
ahaller2: are these examples too complex? need v simple examples at beginning
… build up to these more complex examples
Action: ahaller2 to create simple SOSA example with smartphone and sensors
<trackbot> Created ACTION-365 - Create simple sosa example with smartphone and sensors [on Armin Haller - due 2017-06-06].
<Kjanowic> I like this idea!
Action: SimonCox to add examples from O&M converted to SOSA/SSN
<trackbot> Created ACTION-366 - Add examples from o&m converted to sosa/ssn [on Simon Cox - due 2017-06-06].
Kjanowic: need variety in samples to capture audience
<Kjanowic> Okay, I see your point.
Everyone seems to agree that multiple examples are better
ahaller2: single example in SOSA, more in SSN?
… need small document in sosa
are we talking about (a) examples listed against definition in ontology, or (b) fully worked instances in document
<Kjanowic> but liked it in the end
mlefranc: in document - at beginning of each section included links to worked examples in later chapter
ahaller2: in ontology document how to use skos:example? one or many?
… multiple examples are helpful in ontology document
… wide review seems to indicate that multiiple examples useful
… all worked examples should be mentioned in ontology annotations
Kjanowic: would be good to ensure that ontology document is balanced, similar number of skos:example for each class
… like mlefranc suggestion that each of us makes extended example using as may classes as possible
… this will verify that we share understanding of the ontology
Action: ahaller SimonCox mlefranc Kjanowic DanhLePhuoc to create one example each that uses as many classes/properties as possible
<trackbot> Error finding 'ahaller'. You can review and register nicknames at <http://www.w3.org/2015/spatial/track/users>.
<Kjanowic> Yes
<Kjanowic> like a scenario
<mlefranc> +1
Action: armin SimonCox mlefranc Kjanowic DanhLePhuoc to create one example each that uses as many classes/properties as possible
<trackbot> Created ACTION-367 - Simoncox mlefranc kjanowic danhlephuoc to create one example each that uses as many classes/properties as possible [on Armin Haller - due 2017-06-06].
ahaller2: Canonical examples around water bodies?
ahaller2: has been closed
<mlefranc> ok with the new tructure
ahaller2: location and forecasting was in own section, too prominent
… but not much content?
… now moved to subsections 6.3 6.4
Kjanowic: location is not part of SSN - shoudl be delegated
… when we didn't have any real horizontal modules
… suggest moving the things which are not SSN components into yet another chapter after the examples
ahaller2: since we haven't actually modeled them as part of SOSA/SSN they need to be moved to annex?
Kjanowic: the other horizontal modules contain axioms
… for forecasting and locations there is no axioms so no module
… rather this is how to link out to other
… alternatively: we could build an actual alignment
… its a scenario or best practice
mlefranc: should not be pushed into annex. New section "modeling issues'
<Kjanowic> I like Maxime's proposal
<Kjanowic> have a modleing section
would 'alignment module' include stub classes for geometry etc
Kjanowic: yes - to clarify that sensors are not geometries, etc
<ahaller2> ISSUE: Location and Forecasting are not modules at the moment, either they need alignments or move to a seperate section
<trackbot> Created ISSUE-203 - Location and forecasting are not modules at the moment, either they need alignments or move to a seperate section. Please complete additional details at <http://www.w3.org/2015/spatial/track/issues/203/edit>.
<Kjanowic> I will have to leave soon so I cannot scribe
Kjanowic: do it for next week
<ahaller2> http://w3c.github.io/sdw/ssn-usage/
ahaller2: about implementation evidence: need to update ssn-usage document
ahaller2: plug new implementation evidence in ssn-usage directly
Action: ahaller2 update ssn-usage with links to working draft classes/properties and add new classes/properties where necessary
<trackbot> Created ACTION-368 - Update ssn-usage with links to working draft classes/properties and add new classes/properties where necessary [on Armin Haller - due 2017-06-06].
mlefranc: also, draw a graph that show the dependencies (existential relations) between terms
<DanhLePhuoc> https://www.w3.org/2015/spatial/wiki/Implicit_Implementations_entailed_by_Axioms
<Kjanowic> Yes, this was a task we discussed sometime ago
DanhLePhuoc: entailment regime make that class system is instantiate whenever someone instantiates sensor (under ssn)
… also with owl:inverseOf
ahaller2: reflect that in ssn-usage document ?
… separate section (table) for the instance that are implemented "behind entailement"
<Kjanowic> afk
ahaller2: @tidoust, please explain the agenda again
<tidoust> Timeline to reach REC
tidoust: worried about candidate recommendation phase
… meant to be 15th may
… still have time to work on the implementation, need to finalize CR
… still need the "exit criteria" in the spec
<tidoust> Exit criteria for Time Ontology
tidoust: also need the whole group to accept to switch to CR
… and how the wide reviews were handled
<ahaller2> https://www.w3.org/2015/spatial/wiki/Wide_Review
<SimonCox> I need to leave now
Action: ahaller write Disposition of Wide Review comments in Wiki, add exit criteria to document and ask Ed for Plenary meeting to get into RecTrack
<trackbot> Error finding 'ahaller'. You can review and register nicknames at <http://www.w3.org/2015/spatial/track/users>.
tidoust: still work on non-normative sections in document while CR
… can work tomorrow and wait for a few days, 2 weeks, for a consensus (list and director)
… but it needs to happen soon
Kjanowic: what about use of MUST BE and SHOULD BE ?
… quantification, in text box, we use MUST BE
… the axioms and the text are not in sync
… only is forall quantifiers, exists is for existential quantifiers, need to check consistence in the figure
<ahaller2> change axiom language, must be -> only, at least one -> exists one
<Kjanowic> \forall --> only
<Kjanowic> \exists --> some
<Kjanowic> cardinality --> exactly 1 or at least 1 or at most 5
<Kjanowic> ....
Kjanowic: *discussing the wording for every axiom in the spec and the figures*
ahaller2: need to be done before vote tomorrow
Action: ahaller to change language for axioms to match best practice
<trackbot> Error finding 'ahaller'. You can review and register nicknames at <http://www.w3.org/2015/spatial/track/users>.
Action: armin to change language for axioms to match best practice
<trackbot> Created ACTION-369 - Change language for axioms to match best practice [on Armin Haller - due 2017-06-06].
<Kjanowic> I have to leave now
mlefranc: why do we use slash-based URIs anyways ?
Action: mlefranc to harmonise hash URIs to slash URIs for horizontal modules
<trackbot> Created ACTION-370 - Harmonise hash uris to slash uris for horizontal modules [on Maxime Lefrançois - due 2017-06-06].
mlefranc: only important if one can get different documents from the same namespace
DanhLePhuoc: plan to demo something next plugfest in July
<ahaller2> ... next plug fest in Duesseldorf
DanhLePhuoc: plan to use SSN for anything relevant to sensor and actuator
ahaller2: can use that for implementation evidence ?
DanhLePhuoc: yes