SDW WG Face to Face, 11 March 2015 -- 11 Mar 2015

<Ioannis_Athanasiadis> Good day!

<Ioannis_Athanasiadis> Seems that the GotoMeeting link for this morning is wrond.

<phila> Yes, Ed is looking into it

<Ioannis_Athanasiadis> Thx :)

<phila> scribe: phila

Intros

kerry: Opens the meeting, runs through the agenda

<jtandy> http://www.w3.org/2015/spatial/wiki/F2f_Barcelona

<jtandy> scribenick: jtandy

<scribe> scribe: Jeremy Tandy

agenda for today: http://www.w3.org/2015/spatial/wiki/F2f_Barcelona

kerry: main job for today is to work through the use cases

<phila> Use Cases

kerry: we're going to organise those use cases and define the requirements
... each use case will be set to the deliverables for the WG
... we finish the meeting with lightning talks

<Alejandro_and_Armin> Same problem with pin code

kerry: introducing the use case editors; Frans K and Alejandro_and_Armin

<Rachel> ditto, problem with access code

phila: does anyone object (in the room) to me taking a picture and tweeting it

(no objections)

kerry: THE MEETING IS OPENED!

[ed returns ... hopefully the goto meeting is fixed?]

kerry: Welcome ... I'm Kerry, co-chair, work for CSIRO in Canberra
... introducing Ed Parsons, fellow co-chair, works for Google in London
... lots of people here - some new faces

<Alejandro_and_Armin> the go to meeting access code is invalid

kerry: lets do some other introductions

<Alejandro_and_Armin> 353923093

[going around the room here ... suggesting that if people want to say who they are - type into the IRC]

scribe: W3C staff is phila; OGC staff is IngoSimonis

<phila> phila: Notes there are 36 people in the room, only 20 on IRC, not all of whom are in the room. We need to fix this

kerry: thanks - now we've got to get to work

phila: if you're not on IRC, you're not on the meeting ... please join [of course if you're not on - you can't read this!]

eparsons: goto meeting is now working

kerry: will now introduce the use case chairs
... Frans (sitting in the corner)
... Alejandro_and_Armin (somewhere in the ether)

<eparsons> Anyone having trouble with Goto Meeting ?

<AZ> Still the same "access code not valid" message

kerry: the main job for today is to organise the use cases into groups / categories & define the requirements from those use cases
... let's kick off

<Alejandro_and_Armin> same problem that AZ with gotomeeting via skype

kerry: we're going to split into subgroups at some point this morning ...

<eparsons> Use this gotomeeting id https://www4.gotomeeting.com/join/182870253

<AZ> BTW, AZ is Antoine Zimmermann

kerry: we have a few ways to organise ... by deliverable (that would make 4 groups)
... alternative to organise by theme
... any ideas?

[discussion - where are we going to put the subgroups]

<Alejandro_and_Armin> ok, we are connected!

<Alejandro_and_Armin> thanks, Ed!

kerry: we could split up by themes ... e.g. IoT, Earth observation,
... socially construed notions of place
... discovery catalogue type use cases
... probably easiest to organise by deliverables

Frans: could allow the sub-groups to self organise?

kerry: the deliverables are:
... 1) (spatial) best practice
... 2) Time
... 3) SSN (semantic sensor network)
... 4) Coverage

[getting straw poll]

jtandy: let's merge SSN and Coverage groups

kerry: OK

chrislittle: not everyone voted

kerry: each sub-group will need to scribe

<phila> If you want to see what the IRC looks like as minutes, see http://www.w3.org/2015/03/11-sdw-minutes.html

phila: you can scribe on paper or electronic form ... you don't necessarily need to use IRC
... PLEASE DON'T USE A PROPRIETARY FORMAT
... when you upload your notes

[discussion - local logistics]

kerry: any questions before we break up

xavier: some of the UC's are very application specific
... some are technology centric
... how should we proceed?

kerry: see the scope questions http://www.w3.org/2015/spatial/wiki/Scope_questions_and_Requirements
... and the charter http://www.w3.org/2015/spatial/charter

eparsons: the four scoping questions are:
... 1) Is the use case specifically about Spatial data on the Web?
... we had a lot of discussions about SPATIAL!
... 2) is the use case including data published, reused, and accessible via Web technologies?
... 3) Has a use case a description that can lead to a testable requirement?
... 4) public vs private web
... all the requirements need to be hung on the use cases ... our requirements need to be grounded in reality

peterbauman: please can we have W3C savvy folks in each group - because most of us are OGC folks?

kerry: don't think we need this ... most important thing is to SCRIBE

eparsons: ditto

kerry: work to get consensus ... but we're not operating a formal decision making process today; discuss, be inclusive
... [notes the PATENT RULE statement]
... talk to IngoSimonis if you've got concerns

<eparsons> Thanks AZ

phila: W3C and OGC have a slightly different method of dealing with IPR ...
... but both horrible processes mean that the standards are royalty free

<phila> The OGC Patent Call

phila: BTW - the standards will be joint badged

kerry: who's going to lead the groups
... best practice ... eparsons
... time ... stefanstrobel
... SSN & Coverage ... [Kerry gets self nominated]

jtandy: when to reconvene back into plenary

kerry: let's get back to do sanity check / share discussions at 12:45 ... just before lunch

cperey: I'll take an action to figure out if the big room will be ready

kerry: we'll meet in the big room at 12:45 (assuming all ok)

eparsons: at Google's great expense I've brought post-it notes ...
... please don't write on the walls :-)

<phila> Meeting breaks into sub groups

<AZ> tables and chairs moving around...

<eparsons> All, Goto Meeting shutting down now until presentations at 3pm CET

<Alejandro_and_Armin> ok, is there a way to join the SSN group remotely?

<Alejandro_and_Armin> thanks!

<Alejandro_and_Armin> my skype id is ale21385

<phila> Charter

<phila> scribe: phila

<scribe> scribeNick: phila

jtandy: Suggests to triage the use cases and classify them into those that are relevant to SSN and Coverage

Much messing with equipment

SSN and Coverage deliverables

kerry: Introduces Alejandro_and_Armin who have joined via Skype
... Talks through the charter
... SSN was created in W3C space but isn't a formal standard.
... widely used
... being asked for help with encoding the data so that might be what this WG does (which will create a formal standard)

-> http://www.w3.org/2015/spatial/charter Charter

<jtandy> http://www.w3.org/2015/spatial/charter

kerry: Concept is coverage over space. Gridded data over space, how you might link to other data
... WaterML2 is mentioned to be able to handle time at a specific point in space
... RDF Data Cube http://www.w3.org/TR/vocab-data-cube/

PeterBauman: OGC already has well defined defn of coverage
... distinction between discrete and continuous understood
... consensus should be based on existing defn of the word

<jtandy> (the charter says: "expressing discrete coverage data conformant to the ISO 19123 abstract model")

jtandy: Points to wording in charter to help clarify what is and isn't in scope
... I suggest we look at the use cases we have.

pebau Distinction between the coverage and the services that operate on it

JohnHerring: Do you know the difference between discrete and continuous
... all data is discrete at its core
... continuous are interpolated, discrete is not

pebau If you're working on an image then you do have pixels. Coverage describes both in a uniform way

kerry: My understanding is that we're working on discrete, not interpolating

pebau Gives an example where interpolation doesn't make sense

JohnHerring: There's no interpolation between and oak and a pine

jtandy: In a gridded daatset - in meteo, we represent each individual point as you need to know when to interpolate and when not to

pebau If it's the case that it can be interpolated then it can be seen as continuous but you won't necessarily provide the interpolation method

JohnHerring: We need to know what you're trying to do before we decide

kerry: Yes we need to look at the UCs!

<jtandy> (the interpolation in meteorology is really quite complicated - so if you're going to do interpolation, proceed with caution!)

kerry: Hands the chair to Jeremy

<scribe> chair: jtandy

jtandy: We're trying to pull out the requirements from each UC. Please read through the first UC
... starting with https://www.w3.org/2015/spatial/wiki/Working_Use_Cases#Meteorological_Data_Rescue_Use_Case_.28Best_Practice.2C_Time.2C_SSN.2C_Coverage.29

Jeremy hands out post-it notes to write the reqs on as we'll have the smae reqs multiple times

kerry: I can see multilingual search as a requirement

Phila: this is now 'mdr' https://www.w3.org/2015/spatial/wiki/Working_Use_Cases#mdr

<Alejandro_and_Armin> are we using any google document or something similar?

Ted: Do we have a quality metric?

Discussion leads to some quality and prov reqs

JohnHerring: If the data is tabular-numeric that may not be stats
... we don't know what the data is

jtandy: These would be reports from a specific station. It's a time series

JohnHerring: One station or multiple stations

jtandy: Each station has its own time series

JohnHerring: So the user might have a consistent set across which she could interpolate

jtandy: It's a continuous phenomenon so if you know what you're doing you can interpolate
... And we normally use quite complicated models to do that
... We have time series records for observations at a point

kerry: WaterML is relevant here then

<stefan> #time The OGC TemporalDWG has already put together a list of useful links wrt time defenitions http://external.opengeospatial.org/twiki_public/TemporalDWG/WebHome#TemporalDWG_Web_Utilities

s/PeterBaumann/pebau/g

Further general discussion about the use case

jtandy: We need to know where the data is related to

pebau: Yes, it must be referenced in space and time

jtandy: Summarises the clusters - when and where
... on the Web we need to be able to find the info - so discovery metadata
... need to be able to cite it
... its quality
... handle multi-lingual
... so in terms of coverage, we're looking at metadata of it

kerry: Need to be able to insert it in a catalogue

Sylvain_Grellet_BRGM: So she's processing the data?

kerry: No, she's just publishing what she finds

<Alejandro_and_Armin> +q on adding requirements to spreadsheet

-> https://www.w3.org/2015/spatial/wiki/Working_Use_Cases#hzv Next USe case

<scribe> chair: kerry

jtandy: Summarises his use case
... based on coastal habitat zones

<Zakim> Alejandro_and_Armin, you wanted to comment on adding requirements to spreadsheet

Looking at https://www.w3.org/2015/spatial/wiki/Working_Use_Cases#hzv

jtandy: We don't want to describe the zone itself, we need line of sight to the info that someone used to designate the protected zone
... so it's really about provenance
... we have a bunch of complex data that may be used
... side scan sonar can be many GB
... Cameras towed behind a boat to show the different plants and fish etc
... person will record what type of fish is seen where

JohnHerring: So we're talking about GeoReferenced video
... Frame by frame georef

jtandy: So all the metadata is in the observation model. Where and when the pic was taken
... The interesting thing for me is, a link to a big side scan sonar dataset - that's too much
... we need to access a small part of the data
... Politicians need to know whether the fish was seen in or outside the zone
... but activists will want more detail
... so we need a provenance record like "I used this chunk"

Ted: Is that a subsetting requirement?
... If I understand the charter correctly - this coverage data is really big so we need ways to share smaller pieces of it
... That's a server side requirement?

jtandy: I was thinking about the citation angle
... they may have had the whole thing

Jeff_Walter: You need to be able to express that you used array indices 60-80
... for example

pebau: We have subsetting in coverage

Discussion focuses on referencing part of a coverage

<stefan> #time agreed on the necessity to define a harmonized OWL model for time and test it with the provided use cases

jtandy: And you, pebau, said you wanted to be able to describe it such that someone else can execute that in future
... I'm going to suggest we break for coffee?

christine: No, 10 mins to go yet

jtandy: There are things in this UC related to BPs and time but we're not looking at those here

JohnHerring: You need to be careful when you say georeference
... every observation has its own location associated with it

pebau: raises the issue of CRSs

Sylvain_Grellet_BRGM: raises identifiers

jtandy: We said in the previous UC that we needed to be able to cite
... and as we want to be able to reference a a specific part of the data

Sylvain_Grellet_BRGM: We use XML IDs for that

Ted: Suggests it's time for coffee

JohnHerring: Historical data is consistent but may not be accurate
... there's a lot of metadata needed to be able to repeat the processing

Discussion dissolves into coffee

BREAK

<Payam> scribe: payam

<scribe> scribenick: payam

<scribe> scribe: payam barnaghi

ManolisKoubarakis: presenting the next use-case (real-time wildfire monitoring)
... the use-case focuses on satellite imaging and the use of ssn
... need for metadata: privacy, ...
... metadata about sensors, satellite sensing

JohnHerring: for this use-case: need to know the geometry of sensors that are used

kerry: we are going to work pst geo-rectified data

s post/post

<laurent-oz> FGDC has worked on this http://www.fgdc.gov/standards/projects/FGDC-standards-projects/csdgm_rs_ex/MetadataRemoteSensingExtens.pdf quite some time ago.

<laurent-oz> ... https://www.fgdc.gov/standards/projects/FGDC-standards-projects/csdgm_rs_ex/remote-sensing-metadata

jtandy: this one is a Geo-Reference, geo-rectified data related to a specific case

kerry: this use-case also discusses publication of raster and vector formats

Li Ping: we have a use-case on this topic in OGC testbed no 5

JohnHerring: there are also other projects that search metadata; Australia, NASA, ... helping technologists to find others who work on a simialar use-case; something like a social network

kerry: this will be out of scope- unless we refer to the description of the method

<Alejandro_and_Armin> shouldn't be this use case also tagged as Coverage?

kerry: a requirement can be: mechanisms to link to social media; making processes sharable and discoverable

JohnHerring: getting people to do crowdsourcing stuff

Payam: this can be a part of the provenance data and linking it to social media; e.g. wildfire

Sylvain_Grellet_BRGM: we use social media for linking the data to social data

<Sylvain_Grellet_BRGM> ex : for seismic events

kymw: this can be used as a use-case to save the wildlife; e.g. describing and finding events, ...

Kerry: this is relevant but not in the scope of the current use-case

<danbri> (fwiw "default" is a really tricky idea formally in RDF/OWL)

kymw: fuzzify location information; e.g. to protect endangered species

JohnHerring: how other data that will be linked to the data in this use-case is going to be discovered?

jtandy: providing hooks to opendata

JohnHerring: how make connection to data that can be shared but it is not public data (do we care if it is open or closed)

Kerry: changing "Open" to Web; open data--> Web data

ManolisKoubarakis: discussing the next use-case: Diachronic burnt scar mapping

similar to the previous usecase

JohnHerring: this is a very useful use-case

jtandy: this use-case can be generalised to other domain such precision agriculture

Kerry: next use-case Locating a thing; this is a bout "time"
... use-case: Harvesting of local search content
... requirment in this use-case: representing spatial and temporal location

also: date, duration, repeating intervals,

next use-case; "Locating a thing"

this will be covered in best practices

Sylvain_Grellet_BRGM: this use-case has a linkage with the OGC IoT group

next use-case (no 9) : Consuming geographical... is related to best practice

Use-case no 10: Enabling publication, discovery, and analysis of spatiotemporal data ....

<Alejandro_and_Armin> +q on temporal granularity and temporal vagueness

Kerry: easy to implement requirement is discussed - related to best practice

jtandy: Nominal temporal references, imprecise time definition

Kerry: granularity

varying time granularity

Use-case no. 11: publishing geospatial reference data

jtandy: this is related to best practice

Use-case 12: Integration of government and utility data to enable....

JohnHerring: they have smart metering and accessing sensors on the web

Kerry: requirement: using SSN as a common framework for cross domain sensor descriptions

<chaals> use case 9

use-case no.13: Using spatial data from the web in GIS systems during emergency operations

<Zakim> Alejandro_and_Armin, you wanted to comment on temporal granularity and temporal vagueness

Alejandro_and_Armin: data and Actuation decisions
... whether we are doing for SSN and coverage and time

Kerry: we rely on the use-cases to derive the requirements

back to use-case no.13: Using spatial data from the web in GIS systems during emergency operations

Kerry: exchanging data about incidents

mostly related to best practices

JohnHerring: could be related to dynamic coverage
... discusses SIM City

jtandy: there is a potential to publish predictive analysis about a fire which could be provided as a coverage dataset;

JohnHerring: cellular automata

use-case 14: publication of air quality data aggregations (SSN, Time)

Alejandro_and_Armin: describes usecase 14

Requirement: aggregated sensor values
... how to use SSN for temporal aggregations and representing addresses

s/Jeanb: JeanB: aggregated space and time/aggregated space and time

Sylvain_Grellet_BRGM: there is an EU standard that does this

jtandy: we have to build based on what we have already agreed

Sylvain_Grellet_BRGM: discusses the EU work related to this
... we use reference vocabularies

<Sylvain_Grellet_BRGM> For UC 14 (AirQuality) : Basically use what has been done in the EU Air Quality Directive (ISO 19100 Feature Model + Observations & Measurements) to generate an ontology

Payam: this is not in the usecases: Requirement: linking an observation to a named event; i.e. abstraction

usecase 15: publication of transport ...

Alejandro_and_Armin: describing usecase 15
... part of this could be related to best practice idea
... e.g. sensors and mobility; e.g. sensor on the bus and then having a sensor at a bus stop when the bus is there

Requirement: mobile sensors

Payam: this is related to dynamic semantics; e.g. location value can change and what could be the best practices to update the location data and what mechanism will trigger it.

JeanB: this is more related to sensor but we are more interested in observation

Alejandro_and_Armin: we still use the sensor as a proxy to obtain the location

KymW: in the OGC, we differentiate between moving features and moving sensors

use-case 16: Combining spatial RDF data for integrated querying ...

related to best practices

use-case 17: Dutch base registry

realted to best practice

use-case 18: we skip this (Kerry's use-cases)

<scribe> ACTION: Kerry to review the list of SSN use-case and come up with the requirements [recorded in http://www.w3.org/2015/03/11-sdw-minutes.html#action01]

<trackbot> Created ACTION-10 - Review the list of ssn use-case and come up with the requirements [on Kerry Taylor - due 2015-03-18].

use-case 19: publishing Cultural Heritage data...

related to best practice

Requirement: to be able to reason over time information (owl reasoners)

jtandy: most owl reasoners do not underestand xsd: time, etc.

pebau: they have worked on this and have a publication

we couldn't see a coverage requirement for use-case 19

Rachel we are discussing your use-case

<jtandy> http://www.w3.org/2015/spatial/wiki/Working_Use_Cases#Dissemination_of_3D_geological_data_.28Best_Practice.2C_Coverage.29

<Rachel> I am on the chat, can answer questions here

<Rachel> example links: http://earthserver.bgs.ac.uk/clients/3d/glasgow_geology.html

<Rachel> example links: http://earthserver.bgs.ac.uk/clients/3d/tno_lithostrat_rgb_voxel_06_cache.html

invitation: please update the wiki and add the references to existing technolgies, papers, et.c associated with the deliverables

Requirements: Publish 3D LIDAR
... publish 3D voxel dataset
... Publish 3D spatial models

Rachel: are you happy with these?

<Rachel> yes, thanks!

<JoshLieberman> phila, can't get on to the phone bridge "restricted at this time". Anything I can do?

Yes Phila

<phila> Your place our ours, Payam ?

your place

<JoshLieberman> I've asked Greg Buehler to start gotomeeting in lieu of Simon

bigger room

<phila> Thanks. We're ready when you are

will you have this Phil

do I need to type anything to save the minutes

<pebau> looks like we are getting into more and more of the coverage types, as defined in the coverage spec (OGC 09-146r2), which is good: one concept (coverage) to address these means we can achieve a homogeneous service landscape as well

<Alejandro_and_Armin> good night, guys! keep up the good work

<pebau> ...see Web Coverage Service (OGC 09-110r4) and http://external.opengeospatial.org/twiki_public/CoveragesDWG/WebHome

Comment to Rachel: you may consider adding Ten - tetra hydro irregular networks - and pixels

Rachel this comments was from JohnHerring

Summaries

<phila> Kerry and Frans summarised the discussions held in the two groups (see their minutes)

<phila> *** LUNCH ***

<eparsons> https://www4.gotomeeting.com/join/682717181

<jtandy> scribe: Jeremy Tandy

<SimonCox> I just started the GoToMeeting

<jtandy> scribenick: jtandy

<SimonCox> Hi JT

<SimonCox> no

<Zakim> Rachel, you wanted to say publish 3D spatial models requirement should include handling 2D or 3D TIN (=TEN tetrahedral network) and heterogenous resolution grids

<SimonCox> which audio channel should I be using?

<Payam> Simon: call you call me on Skype

<Payam> can?

<Rachel> the publish 3D spatial models requirement should include handling 2D or 3D TIN (=TEN tetrahedral network) and heterogenous resolution grids

<kerry> mattperry, gottomeetning is in a diferent room to irc

<kerry> this room is time, ssn, coverage

<SimonCox> DO I leave the GoToMeeting then, and reassign my mike and speakers to some other device?

<kerry> if you want best practive, gotoemeeting should work

<Payam> yes Simon; can you use Skype

<Payam> ?

<kerry> irc people, thie irc is doing ssn, coverage and time

<kerry> icr peopole, if you want best practice, fo to gotomeeting

<kerry> at 4pm we will all join up, irc and gotomeeting will be in same place

<kerry> 4pm!

more discussions of use cases (for Time, SSN & coverage)

use case: Publication of Raw Subsurface Monitoring Data

http://www.w3.org/2015/spatial/wiki/Working_Use_Cases#Publication_of_Raw_Subsurface_Monitoring_Data_.28Best_Practice.2C_SSN.2C_Time.29

kerry: just to remind people that we're working to the deliverables in the Charter @ http://www.w3.org/2015/spatial/charter#deliverables

<Payam> Rachel: are you there?

<Rachel> yes, I'm here

kerry: Rachel - can you suggest what your key requirements are for this use case?

<Rachel> best practise on how to publish time series monitoring data so that it can be consumed by clients over the web

<phila> Publication of Raw Subsurface Monitoring Data Use Case

Payam: I think the requirement here is to annotate the streams & link to the observation [...]

<SimonCox> 3-D, time-series

kerry: 3d, location, feature of interest with Time Series measurements

<SimonCox> Also: 3-D location may not be in x-y-z - borehole data is offset along a curve

<chaals> where we are up to

Aaron: (geoscience australia) says that Rachel's use case is very similar to theirs

<SimonCox> Also: 3-D location may not be in x-y-z - borehole data is offset along a curve

<Payam> Requirement-> Rachel: practise on how to publish time series monitoring data so that it can be consumed by clients over the web

jtandy: calls out the point about boreholes

<SimonCox> Yes - local frame

liangsteve: this is a local frame of reference

kerry: is this our requirement or best practice

jtandy: best practice

kerry: [moving to the next use case]

(oops - not moved on yet)

kerry: requirement here around timeseries graphs
... low accuracy is not an issue for us; we already talked about data quality and imprecise references for time
... [adding another requirement for imprecise spatial references]
... imprecise location
... any more requirements in this use case?

aaron: thinks that this is covered

https://www.w3.org/2015/spatial/wiki/Working_Use_Cases#placeNameOnt

next use case: Use of a place name ontology for geo-parsing text and geo-enabling searches

<SimonCox> SIRF project at CSIRO

SimonCox: am reading carefully what Rachel wrote

kerry: looking just at the times

Sylvain_Grellet_BRGM: this raises issues about geological time

SimonCox: some of the key things here are changes of name over time

<Sylvain_Grellet_BRGM> this was discussed during the "time" split session this morning. They added new use cases

kerry: ok - so naming places is best practices?
... the other group's job

phila: potentially ... this use case suggests creating a new ontology

<arogers> also, the extents of a named place changing over time

phila: be surprised if we need to do this

<SimonCox> THe main requirement relates to geo-parsing

<SimonCox> I agree that time is a minor element of this story

JeanB: this also picks out the need for explicitly defining temporal reference system

kerry: ok - new requirement

<SimonCox> Versioned names is good

phila: Rachel is suggesting that geonames _already_ has versioned names

Driving to work in the snow

<phila> Driving to work in the snow

<Sylvain_Grellet_BRGM> @Simon : but we don't usually refer to reference objects using their name (ex : a monitoring facility, a borehole, a geol unit)

Payam: this is talking about point of interest
... this is about linking the observation (event) to the real world feature / thing

kerry: and the real world feature has spatial attributes

JeanB: we're also interested in WHEN the observation happened

Payam: this also talks about lightweight protocol ... integration with MQTT etc. for IoT

liangsteve: [agrees]

kerry: so the requirement is "lightweight version of the ontology for IoT applications"

Payam: we have a requirement to provide tools and APIs to help people publish / discover this content
... like a reference implementation

phila: nothing in W3C becomes a standard without demonstrating that it works
... you need two independent implementations

jtandy: so we need a reference implementation to complete the REC (recommendation) track

<SimonCox> OGC has test-suite in principle, but often not available until too late ...

phila: W3C also encourages use of test suites
... html5 has thousands
... with SSN (etc) you're validating the data - not testing the sensor

<SimonCox> SPARQL assertions for validation

phila: the basic thing (for the test suite) is to prove that it works

<SimonCox> Cory is now online

<Payam> we have developed a simple validator for SSN: http://iot3.ee.surrey.ac.uk/SSNValidation/

<Payam> CoryHenson: we are discussing your use-case

<Payam> https://www.w3.org/2015/spatial/wiki/Working_Use_Cases#snow

kerry: requirement - need to represent qualitative observations

<SimonCox> (i.e. CategoryObservation from O&M)

liangsteve: trying to match this to O&M
... O&M has category observation

<Zakim> jtandy, you wanted to point out what that CSV on the Web is building a test framework

ack

<CoryHenson> Payam: Is anyone else on GoToMeeting besides myself and Adila?

<Payam> no Cory; can you join on skype?

<SimonCox> At least one test per requirement

<Payam> GoTo meeting is in a different room

jtandy: CSV on the Web is building a test framework ... happy to share the approach with this WG

kerry: ok - moving on ...

<SimonCox> Have we moved on?

kerry: we're talking about spatial relations that are contextual; e.g. "near the coffee shop"
... but this is for the spatial best practice team

<SimonCox> +1

kerry: [...]

phila: this "integration of SSN with existing, well-known, IoT protocols (CoAP, MQTT, etc.)" is just what we need to do

<SimonCox> Who is AZ?

kerry: another requirement here about repeating intervals (schedules), calendars, temporal relations
... I think we're done with this use case

Intelligent Transportation System (Best practices, Time, SSN, Coverage)

<phila> ntelligent Transportation System

<SimonCox> Kerry is using numbers but they are not inline in the list of ucs

<CoryHenson> Payam: How to dial-in with skype? ID: henson.andrew ... I cannot hear anything over GoToMeeting

AZ: in this use case we focus on the traveller side

<SimonCox> (I only have my laptop screen and already too many windows)

AZ: to help the user find the best route
... ... also dynamic information - like the current location of buses / trains, traffic info, public works

<Payam> CoryHenson: sent you an invitation

AZ: geo-information and dynamic information too
... possibly the sensors could be in the road, sensing traffic counts
... assessing traffic flow

kerry: this fits with the earlier discussion about categorical observations
... "something has happened"
... e.g. traffic got heavy
... qualitative observations

<scribe> ... new requirement here: sensors creating qualitative events

JeanB: there might be some interest in [OGC] MovingFeatures here
... also linear referencing; e.g. from junction X, 100 yds

<SimonCox> linear referencing is same as borehole

kerry: still working on "intelligent transport systems" uc

JeanB: the kind of event that could happen along a road network - is this something that should be identified in this case?

kerry: asks for clarification

<SimonCox> [I can create another GoToMeeting if that would be preferred]

JeanB: [...]

<SimonCox> [sound on skype is good, but not everyone can join themselves]

kerry: I think we already have this for the qualitative observations
... we're not going to [build an ontology] for transport networks

liangsteve: we mention spatitemporal scales

kerry: yes - putting data together from multiple scales
... being about to adapt to varying scales

liangsteve: observations can happen at different scales

kerry: requirement - represent and integrate observations from different scales

AZ: parking spots - we need to know when a parking spot becomes available

kerry: already have this with sensors detecting qualitative events?

phila: related; not the same thing ... parking spot being free (or not) is a query of state, I don't care when it became available
... [...]

<Sylvain_Grellet_BRGM> isn't that a kind of notification system ?

Payam: we need an interface to find the data and then query

kerry: out of scope

<phila> phila: Just thinking aloud. Do we want to think about defining a browser API for finding the state of something sensed by a sensor, e.g. whether a parking space is occupied or not.

Payam: do we go to the interface [API] level

<phila> There's a difference between detecting the event that state has changed (activity streams?) and querying its current state.

liangsteve: yes - in SWE4IoT we provide an API; not just the data model

kerry: ok - this is a metaquestion to come back to
... "shall we define APIs to work with this data - or just the data models?"

phila: of course - we need to be careful what we wish for - because we'd need browsers to adopt the API in order to complete REC track

kerry: next use case

Optimizing energy consumption, production, sales and purchases in Smart Grids

http://www.w3.org/2015/spatial/wiki/Working_Use_Cases#Optimizing_energy_consumption.2C_production.2C_sales_and_purchases_in_Smart_Grids

<phila> Optimizing energy consumption, production, sales and purchases

<phila> ^^ The use case we're talking about

AZ: [talks about the use case ... refer to the text on the wiki]
... need both predictive information (e.g. weather) and historic information (statistical analysis)

kerry: requirement - temporal needs to work with past, present and future

<chaals> • Agreed-upon vocabulary for metadata about spatial datasets

<chaals> • Agreed-upon URIs for common spatial reference systems

<chaals> • Recommendations for a default, canonical spatial reference system for spatial data published on the web

<chaals> • Recommendations for serializing geometries as RDF

kerry: ... and there's another one

AZ: the current location of the electric vehicle is important too - they might _be_ sources of energy too

(like a big battery)

kerry: this is like the location of a moving sensor

<Zakim> phila, you wanted to ask Antoine whether TV schedules, sport schedules etc are important for this Use Case?

phila: @AZ - in your use case, does it matter about the TV schedule?
... when a commercial break occurs on TV, everyone makes a cup of coffee - puts the kettle on

AZ: yes ... thanks

kerry: adding this to an existing close requirment
... schedules are already discussed in a previous use case
... moving onto next use case

<phila> 101 Smart City Use-cases

"101" Smart City Use-cases

Payam: in smart cities, we annotate _data streams_ not individual observations

kerry: new requirement - annotate data streams; link data streams to the observations metadata

Payam: another concern is "the human as a sensor" - with observations coming from social media
... we also need to link observations to domain specific ontologies

kerry: linking this to the existing req about qualitative observations

Payam: describing the context of the device that provides the data

liangsteve: 2 parts here; one is the specification of the sensor
... other part is environmental; e.g. the electric vehicle when going down hill _generates_ electricty

<CoryHenson> +1 for linking context to device/observation

liangsteve

liangsteve: linking the sensor to some other "coverage" (like a DEM; digital elevation model) to determine context

Payam: [...] licensing of data

kerry: is this out of scope?

phila: think if another WG can cover this - e.g. DWBP (data on the web best practice)
... push the requirement to them?

<phila> phila: It's OK to find requirements that you can pass to other WGs. For example, Digital Rights Statements, licensing etc. is, perhaps, something that could be passed to the Data on the Web Best practices WG

<phila> ... (or just to me)

<SimonCox> @phila Are DWBP using ODRL?

Payam: another one - data publishers define relationships between data streams

kerry: thanks - adding this requirement

<phila> No, SimonCox - but I'm gathering evidence that we need a full WG to make it a standard. Problem, as ever, is critical mass of W3C Members

kerry: @Payam - is that the full 101?

Payam: yes ...

liangsteve

liangsteve: need to think about participatory and opportunistic data capture
... e.g. "I saw this [and reported it]", vs. "I carry a cell phone that measures barometric pressure"

Payam: when working with large data ...
... do we need validation tools so that people can test compliance against our ontologies / recommendations?

kerry: another meta requirement ... can we come back to that later?
... @phila - can you remind me

<scribe> ACTION: phila to remind kerry about providing validation test suite [recorded in http://www.w3.org/2015/03/11-sdw-minutes.html#action02]

<trackbot> Created ACTION-11 - Remind kerry about providing validation test suite [on Phil Archer - due 2015-03-18].

kerry: next use case

<phila> Linked Data for Tax Assessment

Linked Data for Tax Assessment

kerry: this one came from the wider community ...
... does this have new requirements - or have we already covered this

phila: looks pretty straight forward;
... e.g. "tell me where someone lives"

kerry: ok - best practice stuff
... moving on

Images, e.g. a Time series of a Water Course

http://www.w3.org/2015/spatial/wiki/Working_Use_Cases#Images.2C_e.g._a_Time_series_of_a_Water_Course_.28Best_Practice.2C_Time.2C_SSN.2C_Coverage.29

<SimonCox> I would like to make sure a uc involving Sampling Features gets on the agenda

<SimonCox> AFAIK this requirement has not been mentioned yet

<phila> Images, e.g. a Time series of a Water Course

kerry: the interesting thing here is that the images change with time

<Linda> link to doc with best practice breakout group minutes: https://docs.google.com/document/m/?id=1oxrsKw2scisIt9fSGe-KhTCr8VimKbhJgXzBBv7iaS0&login=1&pli=1

<phila> Thanks Linda

kerry: a coverage could define [...]

[waiting to summarise discussion]

liangsteve: in remote sensing the location of the sensed thing is different to the location of the sensor
... this is the definition of "remote sensing"

JeanB: I think that this is about relating coverages together [...]
... it would be interesting for prediction; e.g. for floods ...

kerry: relating timeseries over time; relating spatial coverages over time

<scribe> ... new requirement

<SimonCox> ... and linked sampling features

<SimonCox> stations on a traverse, etc

Sylvain_Grellet_BRGM: and those timeseries coverages are of images in this use case

JeanB: need to think about the granularity of data that we're annotating with metadata?

<eparsons> It's 4pm CET - Time to Summarise requirements so far in the big room

jtandy: yes - this is a perma thread ... do I provide metadata for an observation with an individual tuple or a 20GB data set ... or some where in between

<SimonCox> An RDA WG has just generated a report on dynamic data citation

<SimonCox> focussed on subsets

kerry: yes; new requirement - "annotation for multiple granularities of data"
... ok - so where going to move in to the plenary session to present a summary of this session

<eparsons> Goto Meeting for Lightning talks is https://www4.gotomeeting.com/join/682717181 at 4:30pm CET

<SimonCox> Which channel now? GoToMeeting? Which one

[all those on skype - we're moving to the other room where the goto meeting is operating ... at least as well as it does]

<SimonCox> Which gotomeeting?

we're going to use the _new_ goto meeting https://www4.gotomeeting.com/join/682717181 set up for the lightning talks ...

<eparsons> Yes audio will be via goto meeting

it is ready right now

Summary Session

<phila> JeanB: We started with the time aspects. We weren't sure where to go with this...

<phila> ... we went through the UCs trying to identify what woujld be needed for the future ontologies

<phila> ... not in tejrms of classes and themes but the more generic sense

<phila> ... how can we establish the ontologies for relationships between coverages, time series, specific reqs etc.

<phila> ... we looked at 28 use cases

<phila> ... the first use cases were more difficult

<phila> ... we found a lot of commonalities

<phila> ... need to go through and create links between use cases and reqs

<phila> kerry: We need to go through the remainder of the use cases

<phila> ... in the telecons

<phila> ... it'll be easier to finish off in the telcos now

<phila> kerry: we had discussions about things that may be going broader, like browser APIs

<phila> JeanB: We know that space, time and moving features are all important

<phila> The requirements post-its

<phila> chaals: We went through a lot of the UCs (24)

<phila> ... some things are very general. We want to be able to find data, query it

<phila> ... what queries can a given service answer

<phila> ... not all stores will implement everything

<phila> ... those are very general to data on the Web

<phila> ... Within that there are specific spatial aspect.

<phila> ... Time is important

<phila> ... nachines should be able to match two spaces

<phila> ... My phone my tell me I'm in Barcelona by giving my location far more accurately than is relevalt

<phila> ... so diff levels of details is important

<phila> ... Shapes are important. Default is WGS84. Doesn't define an actual shape

<phila> ... smart people use a different CRS

<phila> ... they'll know which lines are accurate and where etc.

<phila> ... we need to allow for named places, eg Barcelona

<phila> ... diff people mean diff things but they overlap

<phila> ... need to be able to deal with fictional places (Starship Enterprise etc)

<phila> ... Got to give spatial data a URI so it's on the Web

<SimonCox> Question to Spatial Best Practices group:

<phila> ... can't rely on a heirarchical structure

<SimonCox> Is place disjoint geometry?

<phila> chaals: Need to be able to mix geographic and non-geographic

<phila> SimonCox: Is place disjoint geometry?

<phila> chaals: It can be

<phila> SimonCox: It's a pretty fundamental aspect of the OGC view of the world

<phila> ... we have features have geometries

<JoshLieberman> geometry is a model, e.g. of place. Not a place itself...

<phila> ... this is a key ontological basis that's not always understood in the Web spatial community where there's more of a passing reference

<phila> chaals: We have to be able to describe a geometry

<phila> And therein lies a topic that is lurking under all this JoshLieberman

<SimonCox> I don't think Chaals got my point

<JoshLieberman> Agreed

<Linda> we discussed it earlier in the bp breakout group and I think it came across Simon.

<MattPerry> Simon: I think you meant to say "is place disjoint from geometry"

<phila> Kerry: Please be back at 16:30 for series of quick talks

<SimonCox> OK - thanks Linda

<phila> == BREAK==

<SimonCox> @MattPerry - yes, typing fast

<SimonCox> Phila caught it in minutes OK

<SimonCox> Everyone mute GoTOMeeting if not speaking

<SimonCox> please

Lightning Talks

<phila> This session is being recorded through GoToMeeting. But please put the link to your slides here (upload them to the wiki if you don't have them online already).

<phila> The Locations and Addresses Community group

<phila> The LOCN vocabulary

<phila> for those unfamiliar with DCAT

<SimonCox> Please ask everyone to load slides here: https://portal.opengeospatial.org/index.php?m=projects&a=view&project_id=82&tab=2&artifact_id=62359

<phila> As well as the wiki page at https://www.w3.org/2015/spatial/wiki/F2f_Barcelona

<danbri> SimonCox, what flavour of OWL (DL, OWL-2, Lite, Full etc.) does OWL Time work for? (sorry if you covered this, audio fuzzy)

<CoryHenson> bye

<Adila> bye

<MattPerry> bye

<Rachel> bye

<SimonCox> bye

<AZ> bye

<phila> trackbot end meeting

- DRAFT -

SDW WG Face to Face, 11 March 2015

11 Mar 2015

Attendees

Contents