See also: IRC log
<kerry> scribe: rob
<kerry> scribenick: roba
scribenick roba
<kerry> proposed: approve minutes https://www.w3.org/2016/08/23-sdwssn-minutes
<SimonCox> +1
+1
<DanhLePhuoc> +1
<kerry> +1
<SimonCox> present?
<ClausStadler> +1
RESOLUTION: minutes accepted https://www.w3.org/2016/08/23-sdwssn-minutes
<SimonCox> who is here?
kerry - simon has put background on list, some folloups
simon: refs AGU paper as per link in email - exploring PROV as an upper ontology?
<SimonCox> https://goo.gl/TKlX1l
<kerry> https://goo.gl/TKlX1l
simon: 'punchine on slide 14 "CorePROV– allalignments"
OED: [MASS NOUN] The action or process of closely observing or monitoring something or someone:
ChrisLittle: where does Event fit - works with Time ontologies
simon: PROV Activity matches 'occurent'
roba: who has looked at this and who has issue/ supports?
kerry: where is om-lite?
simon: see slides 4,5 - not
updated to SOSA work
... SOSA had Activity as super-class - people unhappy with
subclassing hierarchy
kerry: use Observing etc?
simon: want to use a Noun - eg ActivityOf
<SimonCox> https://www.w3.org/2015/spatial/wiki/index.php?title=SOSA_Ontology&oldid=2342
ChrisLittle: forecasting under procedure?
simon: ObservingProcedure -> EstimatingProcedure ?
roba: what is the entailment regime supported by core (RDFS? if not then SKOS?)
kerry: RDFS was originally agreed - Krystof backed off - inverseProperties snuck back in.
simon: as per OED example should choose terms that match natural language as far as possible
kerry: concerns that SSNX proposal adds extra complexity
simon: BP work - discussion on
HTTP range-14 - is this related - ssn:observation is an
"information object"
... re rules forn Kerrys paper - support and back up idea of
rule - mutiple paths - this is the "art" - make useful
paths
kerry: fill in missing bits using the rules - more reasoning on the back end
simon: reflects on O&M
... terms matched broad community
roba: modularity and encapusalation preferred way to handle complexity - want modules with specific entailment rules
DanhLePhuoc: performance is an issue - sensors generate a lot of data - cant afford to force 3-store to do a lot of work
<ChrisLittle> S/encapusalation/encapsulation/
DanhLePhuoc: focus somewhere
between RDFS and OWL-EL perhaps?
... problems database trying to solve now
... need to address logic load in parallel with design
... commen on Simon's paper - concerned it requires OWL-DL -
not well supported from performance
simon: on modularisation - should the core be the supr-classes, the most common use subclasses (hit the 80%)
kerry: my concern is what goes in
core.
... what is it supposed to be for
simon: armin and jano may have different ideas
kerry: core is what IoT ,
schema.org etc
... needs?
<SimonCox> Apologies for jumping the queue several times. Wanted to catch the discussion, but bad process, sorry
roba: wants core to introduce common terms, superclasses and establish patterns - and support RDFS entailment explicitly.
DanhLePhuoc: supports explicit entailment regime for core - RDFS - maybe extend to OWL-QL
roba: whats the plan here?
kerry: wanted to air the issues - cant resolve yet - need to bear in mind.
<SimonCox> Two issues on the table:
<SimonCox> 1. Observation - record or activity
<SimonCox> 2. SOSA-Core goals
<SimonCox> Options for SOSA-core: superclasses, vs. most-used classes (for IoT)
Kerry: moving on F2F at TPAC - describes agenda - proposes trying to put resolutions re core design - contribute via email.
<SimonCox> Implications: entailment regime?
<SimonCox> Which day is the meeting at TPAC?
<SimonCox> @Chris 'RecordOfObservation' vs 'ActivityOfObserving'
<kerry> +1+1
<kerry> q/
<ChrisLittle> +1 to Simon's terminology
kerry: something concrete on modularisation would be worthwhile if possible