SDW WG Weekly

Is IRC functioning?

<eparsons> YY

Its prob ably my turn

<eparsons> scribe: simoncox

Approve Minutes

<eparsons> http://www.w3.org/2015/07/08-sdw-minutes.html

<Payam> +1

<jtandy> +1 (approved)

<eparsons> PROPOSED: Accept last weeks minutes

<Alejandro_Llaves> +1

<eparsons> RESOLVED: Accept last week's minutes

<ahaller2> wasn't present

Patent Call

<eparsons> https://www.w3.org/2015/spatial/wiki/Patent_Call

No objections - 2015-07-08 minutes approved

Use Cases and Requirements: ISSUE 13

<eparsons> http://www.w3.org/2015/spatial/track/issues/13

eparsons: Issue013

Alejandro: ISSUE 13 Profiling
... Profiles of SSN 1. constrained model 2. compliance - unclear which?
... understands need to check data is compliant with SSN model - no clear way to do this - W3C RDF Data Shapes probably relevant but incomplete

<Payam> forgot how to add myself to the qeue

<eparsons> "q+"

Alejandro: e.g. geology wants to define version of SSN with specific constraints on values - probably not possible in SDW - must be delegated to application community?

Payam, Chris Little, Armin on Q

<Payam> http://iot.ee.surrey.ac.uk/SSNValidation/

Payam: validation is needed in Requirements

Chris is a chipmunk

<Alejandro_Llaves> helium?

Come down Chris - all forgiven

General hilarity

Armin: 1. RDF Shapes not viable solution 2. different modules of SSNO makes it difficult to define generic validation service

Chris: if SSNO is complex, profiles are essential; if SSNO is simple, profiles implies SSNO is inadequate - which?

Jeremy: SSNO is complex; typically necessary to add domain specific aspects in a profile

<Alejandro_Llaves> +q

Jeremy: RDF Data Shapes is unlikely to be finished in time

<ChrisLittle> +1 jeremy

Jeremy: Is simplifying a complex model for a domain application a 'best practice' in its own right?

Alejandro: do we agree SSNO validator required?

<ahaller2> +1 profile

<ahaller2> -1 validator

Alejandro: do we need SSNO profiles?

Jeremy: is the validator/profile requirement specific to SSNO? Or is this a generic requirement - to be able to profile/validate against data models?

Alejandro: focussing on what goes in document
... set 'solutions' aside at this time?

Armin: what does validator actually validate?

<Payam> +q

Payam: validation allows combination of more than one ontology

Jeremy: 1. validation = verify that data is complete, to support application
... 2. validation = verify that profile is conformant to general case

Andreas: OWL models/ontologies are concerned with logical consistency, not integrity
... RDF data shapes - add integrity checks; QB includes SPARQL ASK queries to check integrity

Ed: not convinced there is big validation requirement

<ahaller2> don't care

Alejandro: Barcelona discussion focussed on validation; requirements on list/document appears to focus more on application-specific profiles

<aharth> link to qb well-formed section: http://www.w3.org/TR/vocab-data-cube/#wf

<ahaller2> it is the web, everyone can extend ontologies how they like

Ed: requirement does not call out validation - can we close issue?

Jeremy: ask validation question in UCR next draft?

<Payam> I'm sorry, I have to leave early today

<eparsons> PROPOSED: Close issue - case for validation not made yet.. will revisit

<AndreaPerego> +1

<ChrisLittle> +1 revisit

<Alejandro_Llaves> +1

<Payam> +1 revisit

Ed: close ISSUE 13 - no case for validation yet (can be reopened later)

<MattPerry> +1

<eparsons> RESOLVED: Close issue - case for validation not made yet.. will revisit

Jeremy: call out 'candidate' and 'deferred' requirements - validation = candidate requirement, not addressed now

<jtandy> Candidate ... Accepted ... Deferred requirements ...

<jtandy> (see http://w3c.github.io/csvw/use-cases-and-requirements/index.html for example)

Jeremy: use precedent from CSV on web

<eparsons> Topic : Best Practice Consolidation Progress

Ed: next - BP til now

<Alejandro_Llaves> I did not

<Alejandro_Llaves> https://www.w3.org/2015/spatial/wiki/BP_Consolidation

Jeremy: has membership reviewed https://www.w3.org/2015/spatial/wiki/BP_Consolidation ?
... propose working through UCs to pull out common themes to use in narrative?

<eparsons> +1

Jeremy: focus is on Spatial Best Practices in general, Time/coverages/SSN only incidentally

<Alejandro_Llaves> +1

<LarsG> +1

Jeremy: publisher vs consumer view - typically publisher wears cost to make consumer's life easier.
... see summary https://www.w3.org/2015/spatial/wiki/BP_Consolidation#Analysis_pointers
... e.g. looking for wildfires using satellite imagery - UC is mostly about classifying pixels; BP can't address details of processing algorithms, but might look at BP relating to inputs and outputs

Ed: yes, separate concerns

Jeremy: workflows out of scope

Andrea: why focus on UCs rather than requirements?
... appears to refine UCR rather than move towards BPs

Jeremy: rationale = arrange BP around narrative stories, i.e. UCs
... will ensure that BP does address real stories
... compress 48 UCs into a small number of narrative stories

<Alejandro_Llaves> sounds good to me!

<AndreaPerego> +1 from me

Jeremy: consolidation and mapping requirements to stories allows us to check completeness

<jtandy> [4.7 Publishing geographical data](http://w3c.github.io/sdw/UseCases/SDWUseCasesAndRequirements.html#PublishingGeographicalData)

Jeremy: BP will not recommend encodings?

Ed: this would be a big gap, risks making the BP not meet expectations?

Ed, Jeremy: provide examples, but not exclusive list - make it clear that other techniques would be possible.

Ed: BP should be as complete as possible; self-contained as far as possible

Chris: BP should include list of formats, with comments on pros and cons of each format

Ed: how long will it take to consolidate themes? How many?

<ChrisLittle> suggest 6 rather than 12 narratives

Jeremy: no more than 12; BP document must be short-enough ... ; 1-11 took 3 hours, 12-48 to go

ANOB

Ed: use discussion tab on https://www.w3.org/2015/spatial/wiki/BP_Consolidation

<AndreaPerego> Around 10 would be reasonable - 6 are probably not enough to cover all the relevant use cases.

Book travel to Sapporo asap

No direct flights to Sapporo

Best prices are via Tokyo

<ChrisLittle> bye(

<Alejandro_Llaves> thanks, bye!

<AndreaPerego> Thanks and bye!

<eparsons> thanks simon !

<LarsG> Thx, bye

<ChrisLittle> bye (squeak, squeak)

<ahaller2> thanks, bye

<MattPerry> bye