W3C

Spatial Data on the Web Interest Group Face to Face Meeting

19 Nov 2019

Agenda

Attendees

Present
Chris_Little, Peter_Rushforth, Rob_Smith, Simon_Cox, Gabriella_Wiersma, Erwin_Folmer, Stephane_Garcia, Christine_Perey, Joost_van_Ulden, Heidi_Vandarys, Michael_Gordon, Hubert_Lepori, Fabien_Benetov, Peter_Parslow, Ted_Guild, Linda_van_den_Brink, ClemensPortele, peterr, MichaelGordon, Kris_McGlinn, Mathias_Bonduel, Joseph_Abhayaratna, Anna_Wagner, Rui_de_Klerk, Maxime_Lefrancois, Esa_Tiainen, Josh_Liberman, George_Percival, Scott_Wilson, Friso_Penninga, Satish_Sankaran, Clara_Boyd, Gerald_Wong
Regrets
Chair
Linda
Scribe
Ted, Clemens

Contents


Linda: welcome everybody. Jeremy sent his regrets as he isn't feeling well so I will be chairing solo
... I see some new faces so will explain how we work in this group, we use W3C process and tools eg WebEx instead of Goto Meeting
... we use irc and encourage everyone to join as that is where we take minutes and will be rotating scribe
... you can also make corrections realtime

<ChrisLittle> s/realtine/realtime/

Introductions

Linda: agenda is partially fixed for remote attendees, other parts fluid so suggestions welcome

Ted:cite> Ted: I work at W3C and have been focused on automotive standards.

Linda: I'm Linda, chair of this group and based in the Netherlands

Peter: I'm with Ordinance Survey

Fabien: I'm with EU Parliament

Chris: UK Met Office and have been part of OGC for number of years, working on a new initiative for environmental data

Hubert: EUROControl. trying to promote the use of technologies in ATM

Mark: also with OS, project for business and government

Heidi: Danish Academy at Data Supply

Joost: focus on linked data and metadata, I'm with Natural Resources of Canada

Christine: independent consultant, focused on augmented reality

Stephane: IGN hydrology and metadata

Erwin: University of Twente and cadastre

Gabriella: I'm a grad student working with Geonovum

[agenda review]

SSN extension

Simon: thank you Linda for reviewing the pull request. the last working draft is about a year old now. Editor's draft points to this version
... this was prompted by some work for a research organizations out of Queensland during vegetation, soil, biodiversity etc surveys
... they needed a data model to combine their findings
... this revealed some gaps and prompted these extensions
... one or more properties are held constant so you can observe variations, could be sensor readings across time, space etc
... OSN ontology did not support this so added ObservationCollection - series of observations
... what is most interesting in the variation
... if you allow ObservationCollections to be nested you allow for more scenarios
... the information variation can be represented multiple ways

https://www.w3.org/TR/vocab-ssn-ext

Simon: I have generated the rdf and turtle files for these. the second set of concerns was desire to represent the featureOfInterest in sampling
... this facilitates data discovery
... this is one additional class and four additional properties. I have built the sparql construct queries that would fill in the gaps for eg ultimateFeatureOfInterest
... anyone familiar with this document will notice it is substantially rewritten
... I want to work more on the alignment section at the end

Linda: I was in the observations and measurements WG and they are interested in this as well
... should you point from individual observation to collection or vice versa?

Simon: doesn't matter in RDF world as they are just pointers. collections are synthetic representations where the Observations more defined

[refers back to document diagram of different representations of a collection]

Simon: I am following that group on Github

Linda: some didn't like the term ultimate in ultimateFeatureOfInterest and may propose an alternate

Simon: interested in the suggestions

Linda: we want to promote this further on W3C REC track

Simon: I would be content with a W3C Note

Linda: as an Interest Group we are not able to produce RECs but Notes

Simon: as far as W3C is concerned is to publish an updated public Working Draft for this document
... as far as I know, it just requires consent from this group

Ted: happy to help with the formalities, also "EverGreen" process is being discussed and appropriate for things like ontologies

Simon: currently in a branch and happy to merge when appropriate
... should I announce this to the group?

Linda: yes and we will merge after the feedback
... thank you Simon

WebVMT

[Web Video Map Tracks]

Rob: I am an Invited Expert in this group, leading the development of WebVMT

@@slides

Rob: it is designed for Web primarily for indexing and sharing. webvmt.org gives an overview, has a blog and some demos including a new one from last OGC meeting in June in Belgium
... I have since edited and made it more interesting

current Editor's Draft

Rob: there was a metadata search breakout session in June, to index video metadata to eg search by location with a margin or error (~10m)
... I added those search use cases in September. you can have remote maintenance of a windfarm based on evaluations from drones
... there is a github issue for data synchronization, arbitrary data in JSON format but intended primarily for GeoJSON
... you could attach to all sorts of sensors
... WICG (Web Incubator Community Group) was proposed at W3C TPAC in Lyon for exposing metadata from browsers
... there has been work done on this in the past but not agreement nor support. interest renewed so hopefully will be supported in browsers
... Chris Needham (sp?) participated in the latest TPAC in Japan and proposed user generated events instead of agent
... Web Player - JS playback engine, demo at webvmt.org
... finally a video capture app, trying to build a community around it at present
... a couple things that have come up DataCue - when page assets load there are global datacues - level of detail in map and when the video plays a timed datacue
... last I wanted to discuss Metadata search. there are two processes going on, the initial indexing that crawls the assets and extracts the metadata
... creating a db
... then there is the searching process, make a request to the search engine and results are offered through the datacute interface
... I am planning to do a study to proove it works and quantify the performance
... there is also an ethical use of location data discussion which may feed into a best practices document
... you may want to protect against unauthorized facial recognition for example
... you want to be able to restrict a crawler trying to index video for facial recogition and block with webvmt
... I'm looking to get work started in the benchmark program. I am trying to track down geotagged video datasets, welcome pointers
... please introduce people you may know who may be interested in this
... there is a github issue on @@ and camera attributes - the gimbal in drones instead of fixed dashcam

Linda: you said you are in the incubation phase and getting some involvement from WICG

Rob: that is making progress

[Ted mulling over who he may know that might be interested]

Fabien: I saw the video you shared earlier but wonder if you have examples for 360 videos (for AR) or if you even tried it yet?

Rob: this is a draft, please look at the existing use cases around 360 video. any specifics?

Fabien: it should be intergratable, not sure if anyone has tried so far. I am starting to think of longshow - videos of long traversal of Norway
... I don't have any examples yet, do you?

Rob: yes, there was something at ethical location event last month. there was one for military uses to investigate a location
... you can convert from other formats to webvmt and bring it online
... it is an enabling technology, allowing you to take existing data and bring it in and applicable to 360 video
... Christine who is there mentioned a VR tutorial use case, evidence of action. used to prove someone followed a defined procedure - data annotated video

Chris: there is a use case we have in evaluating cloud observations

[power strips arrive in room to applause]

<utopiah> ChrisLittle: thanks, great example, any link to it please?

<ChrisLittle> Rob: will look for public link to 360 deg cloud observing training videos

MapML

Peter: SVG and other groups have joined to help with Web Mapping, this has turned out to be a rather big job
... there are several existing popular implementations
... we have fundamental use cases already broadly implemented
... others being documented in realtime, will find link and post
... issues are being used to gather feedback on that document
... there is a bit of something for everyone and may be worth a review by this group
... we had a f2f meeting at TPAC in Japan in September, attended by some browser and adjacent browser developers
... latter is helpful as it brings in more (Agalia and Beaucoup) than just the handful of browser vendors
... Firefox, Chrome were part of the discussion as well. we need to get a good community to support mapping on the web
... in OGC world there is the testbed program (currently #15, wrapping up) including vector tile parsing with interesting results
... MapML is an HTML extension that would try to capture the spirit/essence of our use cases and requirements
... this year's work was focused on vector encoding, representing properties in html
... other coordinate based features are in JSON and using CSS selector model to style parts
... hopefully some of this work will continue in testbed 16
... we also talked with W3C about the possibility of holding a workshop on maps on html
... curious if Ted can help with that

<peterr> https://maps4html.github.io/HTML-Map-Element-UseCases-Requirements/

http://www.w3.org/2003/08/Workshops/}

Ted provides some generalities on W3C workshops, strategy and suggest 1:1 call with Peter

<ChrisLit> Link to 360 video cloud trainaing video http://archived.informaticslab.co.uk/projects/skyview.html

[lunch]

Discussion on proposed charter for new period

https://github.com/w3c/sdw/blob/gh-pages/rechartering.md

brinkwoman: W3C groups have an end date
... ... the group has been useful joint W3C and OGC discussions
... ... usually there has been a single person driving each work item
... ... the group has mainly been driving standardization work, but not really worked as a group
... ... discussed the proposal with the co-chair Jeremy Tandy
... ... we think the group should be extended
... ... (goes through the list of items in "SDW-IG will" of the rechartering document)

PeterParslow: There is a lot of work in OGC leveraging Web standards, but these are run within OGC only, because they are simply using the standards, no coordination is required
... ... do we need to add something to clarify the boundary

brinkwoman: Yes, the group has more a coordination role at this point. There might be options for collaboration in the future, Maps for HTML, GeoSPARQL, etc.

peterr: It would be good, if the group would help to engage other W3C members to participate in the workshop

brinkwoman: Next item is a coordinating role between OAB and TAG and maybe ask them for feedback on draft standards from the other organisation

brinkwoman: ... also review OGC IP idea tracker

MichaelGordon: not everything ends up in the tracker, maybe talk to the OGC IP staff directly

brinkwoman: Next is a periodic review of the OGC tech trends
... W3C has something similar

ted: the next item is about community groups, which may be a good option for example for the WebVMT work

brinkwoman: Yes, this group is an early forum until other fora are established in W3C or OGC
... next is about tracking progress of work items in the W3C strategy funnel (this items needs to be reworded to be clearer)

RobSmith: Where would you see WebVMT's home, if not in the SDWIG?

brinkwoman: It clearly has a home here, but at some point a group that can publish new standards needs to be formed

RobSmith: So what does that mean for WebVMT?

ted: The SDWIG could publish a Note to capture the status, to inform the W3C and OGC membership and continue the work in a group that can publish a standard

RobSmith: Thanks, will take the conversation about next steps offline

brinkwoman: Will talk about the strategy funnel later

brinkwoman: ... next is support for developing working group charter

brinkwoman: ... also publishing errata to existing documents

MichaelGordon: Is, for example, the work done by Simon on SSN (not exactly errata) covered by the list?

brinkwoman: Yes, we will update the charter to cover the existing work

HeidiV: The W3C DXWG is working on profiles, OGC does that, too. Should we add something here?

brinkwoman: The focus is geospatial aspects, but profiles is probably a generic topic and not geospatial

ted: Maybe we can help to identify relevant interest in OGC and W3C when charters are written

brinkwoman: But joining groups like the DXWG for non-W3C members would not be possible

ted: Yes, the standard rules for W3C working groups would apply

ted: ... everyone can create issues or provide pull requests

brinkwoman: The SDWIG can play a role in this, contact the chairs if there is a need and we can see how we can help

brinkwoman: Ted, could you tell us more about "living standards"?

ted: W3C has a formal process document and is introducing the idea of "evergreen standards" so that they can continually evolve

ted: ... this is especially important for ontologies

ted: ... still unclear how this works with time-limited charters, W3C working on this

ted: ... this IG is in a way doing this already

brinkwoman: Could we also have no end date?

<ted> https://www.w3.org/wiki/Evergreen_Standards

ted: This is probably going to far, but the rules are going to be debated for a while

<peterr> +1

<RobSmith> +1

brinkwoman: Is there general support for progressing this?

<ClemensP_> (nodding in the room)

<ChrisLit> +1 to progressing

ted: Are there other topics for collaboration beside the ones discussed?

ted: ... probably "profiles" that was mentioned

HubertLepori: Would DCAT also be a topic? Aviation domain is looking at DCAT and W3C is working on DCAT 2

MichaelGordon: Could we also inform W3C members about the work on OGC APIs?

ted: Next TPAC is in Vancouver

ted: ... maybe it could be presented in the plenary, we could try to get it on the agenda

PeterParslow: Should we be explicit about the list of work items in the charter, like WebVMT, to give a more concrete about the work of the IG

brinkwoman: Yes, we can do that

Discussion with the Linked Building Data Community Group

https://github.com/w3c/sdw/issues/1143

<MathiasBonduel> https://docs.google.com/presentation/d/1bvebnuH_uPAIt9LROcL688LWTco97-jHV73zygaiw5U/edit?usp=sharing

MathiasBonduel: (presents the slides)

MathiasBonduel: ... CG is a diverse group with 20-30 active participants, supports a yearly workshop and has liaisons with several other groups/initiaives

MathiasBonduel: ... collects use cases, requirements, develops best practices and ontologies, in the future evolve to a W3C WG

AnnaWagner: (explains the Building Topology Ontology, BOT, and the Building Product Ontology, BPO)

AnnaWagner: ... open how this relates to geometry

MathiasBonduel: Many of the use cases relate to geometry

MathiasBonduel: ... a wide range of geometry representations are required with different semantics

MathiasBonduel: ... more requirements for processing and querying, 3D is importaant, parametric geometries are needed

MathiasBonduel: ... simple and modular ontologies are preferred

MathiasBonduel: ... analysis of different RDF-based geometries / ontologies has been done

AnnaWagner: (explains OpenCASCADE Ontology)

MathiasBonduel: Different options for geometry representations and for linking exist

MathiasBonduel: ... overall goal is to have a domain-agnostic mechanism of linking geometry in a Semantic Web context

MathiasBonduel: ... text encodings (WKT, GML), binary as text, URI to reference a file

AnnaWagner: (explains OMG levels, OMG = Ontology for Managing Geometry)

MathiasBonduel: (explains FOG = File Ontology for Geometry formats)

MathiasBonduel: Geometry is probably out-of-scope for the CG since it has a much broader scope

MathiasBonduel: ... opportunities for collaboration: Use OMG/FOG/GOM as input to standardization?

MathiasBonduel: ... Linking geometry seems to have overlap with OWL Space proposal

MathiasBonduel: ... Spatial querying in SPARQL - support for different geometry schemas

MathiasBonduel: ... Include use cases and people from other W3C groups?

peterr: (points out relationship with Maps for HTML work and potential exchange of ideas)

brinkwoman: Very interesting work, clearly related to the OWL Space / GeoSPARQL 2.0 work, currently discussed in the OGC Geosemantics working group.

brinkwoman: ... Maybe get involved in these discussions, there is a meeting tomorrow at 14:00 CET

brinkwoman: ... maybe there could be a joint meeting to discuss this in more depth

MathiasBonduel: That is already good feedback, this could be a step in the right direction

josephabhayaratna: The Geosemantics DWG is working on a White Paper to document the current state and describe where we want to go. The work presented would be valuable input to that paper.

MathiasBonduel, AnnaWagner: We can join the call tomorrow

brinkwoman: I will provide the connection details

MathiasBonduel: There is a concern that things can get complex very quickly, but the goal is to keep it simple enough for implementation. Finding the right balance...

<Zakim> ted, you wanted to mention indoorGML and ask about building service extensions

ted: an observation and a question - we talked before; now the new team contact. I wanted to make the connection to IndoorGML for indoor navigation. Intelligent Transportation Systems and Smart Cities are also potential domains to involve.

<ted> i'm coming more from transportation needs and wondering if you have heard of anyone extending your modular approach for transportation needs eg smartcities or its

Structuring semantic feature search

<brinkwoman> solving audio problem...

<brinkwoman> we are reconnecting, but wifi is really slow at the moment

<josephabhayaratna> Unfortunately I have to go. It's nearing 130am here, and I'm fading!

<josephabhayaratna> I'll see (more likely hear) those that are coming to the Geosemantics DWG tomorrow

ogc technology trends

George: interest group gets involved, when an item is a trend what warrants action we create a roadmap
... you've heard future directions and about testbeds

<brinkwoman> Slides for OGC technology trends: https://portal.opengeospatial.org/files/?artifact_id=91143

<brinkwoman> Esa Tiainens slides: https://portal.opengeospatial.org/files/?artifact_id=91212

Esa describes linkage on slide 4

[slide 5 current stage]

Esa: we have low hanging fruit from Nordic mapping agencies
... report on implementation on how it can be possible
... providing use and access to data on web
... idea is to provide successive stages [slide 6]
... benefits can be verified
... we found this is the only way, as you cannot get funding for linking data topic
... need to provide benefits

[slide 7]

Esa: this explains semantics, rdf triple store
... entity in service interface are provided by uris and how to access the store for the content
... it is possible to setup user/ux and search for features by concepts, attributes or attribute values
... we have tested this and it works
... we drop search terms that don't match
... user can select reseources they want to use
... you can provide error descriptions
... we also have elastic search indexing and use that on the backend while the front looks the same

[slide 8]

Esa: problem was granularity was not precise enough
... fragmenting as well, we switched to a more simple approach
... what we had to do for the UX was make an alphabetical set of terms in a drop down box
... usability issues but maximized results
... one idea was to use enriched and transformed ontology like GEMET

[slide 9]

Esa: we are in a proof of concept stage
... what we need for deployment is a simple and easy guidelines for using different tools like shacl, prov-o and shex
... there seems to be many unclear tools that have different aggregations, need a clear roadmap to solve
... need access to metadata, different metadata for different uses
... it can be carried out by experiments but need clear guidelines with tools. hackathons could help move it forward
... at the moment we have a considerable number of tools and EU SDR
... it would be possible to use a graph representation to make this easier to understand the content
... possible also to make a hypergraph
... it can be organized into individual sub collections

[slide 10]

Esa: data contents can be uploaded on a web connection from statics Finland and NLS (national land survey of Finland)

[slide 13]

Esa: process is complicated, called IGALOD
... we can also connect non-spatial data such as census data
... workers are allowed to use data based on their tasks
... end of slides are some examples from @@x project, funding from government is ending
... that is about the whole picture. I want to go back to slide number 9
... guidelines for these tools would be helpful

Ted: perhaps you can form some of your impressions on which tools are useful for what purposes, questions and after they are organized can reach out to the tool developers to see if they can help provide some guidance
... I know some of the developers involved in them but not an expert myself

Esa: we want to provide unique set of recommended tools and consolidate

ogc tech trends

Slides for OGC technology trends: https://portal.opengeospatial.org/files/?artifact_id=91143

George: making data on the web discoverable. what is new or emerging, some coming from Graph workshop W3C put on
... there was support for adding temporal to graph query languages
... there were two new works to come out of JTC1 declaritive propery graph language for SQL and GQL
... SQL/PGQ and GQL relate to one another with some overlap on property graphs
... Tobias from Neo4j will be attending tomorrow to tell us more
... two big questions, should there be different trends coming from spatial data on the web (slide 1? top right)
... we can look at the funnel to see if there is anything new
... there has been quite a bit of discussion around building models. there are other trends around 3d models and like but lack linked data

https://github.com/w3c/strategy/projects/2?card_filter_query=label%3Ageospatial

<brinkwoman> https://github.com/opengeospatial/OGC-Technology-Trends

AR

Christine: they were not interested in spatial data (Asure Spatial Anchors) use cases back in June but wanted to further MS halolens before bringing it to the W3C Immersive Web group
... SLAM was part of indoor mapping pilot over the last 12 months. taking something from point cloud for realtime use
... I see a high need and utility for coordination with Immersive Web group, emphasis was more on VR of late
... AR discussions reopened a couple months ago after VR. we are seeing implementations in Chrome, delays in FF and elsewhere
... there are other proprietary projects for anchors in the real world compared to what MS could do with their ASA
... any device needing to know its location on x,y,z axis, pitch, yaw and roll plus temporal has shared needs (including autonomous vehicles
... also pertinent is scale - planetary, inside a building...

George: may be worth reaching out to Geolocation WG

Ted: more for semi-location for browser use cases - where in the world is the user but not the degree of precision we need by any means
... my predecessor in this group, Francois, is involved in AR/VR etc whereas I am not

Linda: there is a geolocation sensor proposal in devices and sensors group

Christine: they are on my liaison list

Linda: we can help with that

Christine: not sure who is on that from our side

Ted: on W3C end that would be Dominique Hazel-Massieux

Christine: I know him well and he knows my bias
... Fabien who was here earlier is very involved in AR as well
... he and Josh had a hackathon this last weekend

Linda: statistics on the Web best practices is within this group, charter will be extended and work is taking place in Semantics Statistics CG
... they could publish a note through SDWIG
... we didn't discuss SDW Best Practices but seem we are lacking people

Rob: for BP there is ethical use of spatial data topic. Michael is picking it up and just want to flag it at this point

Linda: there are some more general bits on ethics in the current document
... it doesn't explain how to be ethical, you are right it can be developed further

<MichaelGordon> Sorry, had to attend another API meeting. On the ethical side I've raised an issue in Github

<MichaelGordon> For other BP stuff - not much movement since the last meeting. We have some minor errata to fix that is waiting on OGC Policy and Procedure update

<RobSmith> https://github.com/w3c/sdw/issues/1137

Rob: I also wanted to Geopose that Christine raised regarding camera attributes and establishing a mapping

<MichaelGordon> for corrigendium on best practices. There are other issues in the BP project but not really anybody to help out on them

Christine: you want cameras to have an implementation of Geopose?

Rob: yes, also zoom

Christine: that could fall into MIPI Alliance scope. they work on specs for components for mobile phones and also in automobiles
... data transfers, queries etc fall to Khronos group
... they have some camera specs and api definition. that work stopped due to lack of participation from camera manufacturers

Rob: WebVMT is a lightweight solution, does not need to be exhaustive but more rudimentary
... knowing where to look is helpful

Christine: for GeoPose you can say cameras should follow six degrees of freedom protocol
... stay tuned...

<RobSmith> Sorry. Missed Linda's parting comment as audio dropped completely

@@h: earlier on ethical topic, nearby is GDPR and we do not have legal expertise present but maybe should for the Best Practices

RobSmith, she suggested you follow Geopose. trusted you would as Christine admits you are better than she at following up

<RobSmith> Thanks

Linda: there are outstanding issues on BP document but nobody taking it up at present
... Michael would welcome the help
... important to keep it current
... practices evolve as should the document

workshop on transportation and routing

<brinkwoman> ted: motivation for workshop was to increase awareness and to involve more standards groups

<brinkwoman> ... we advertised sdwbp, ssn

<brinkwoman> ... short takeaway is that there should be joint work, working on liaisons.

<brinkwoman> ... mix of academia, standards, industry (eg uber)

<brinkwoman> ... trying to reduce the nr of ontologies that are out there

<brinkwoman> ... more on this tomorrow in the Transportation ad hoc

<brinkwoman> ... top priority is to have an ontology on routing

<brinkwoman> .. something like an observations ontology on top

Ted: session tomorrow afternoon on this but that is the short version

[discussion on terminology for engaging others, avoid ontology, semantic web and even data model. use vocabulary]

[adjourned]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/12/06 12:59:49 $