W3C

– DRAFT –
Spatial Data on the Web IG F2F - Day 1/2

24 June 2019

Attendees

Present
billroberts, brinkwoman, Chrislittle, ChrisLittle_, Christian_Elfers, ClemensPortele, Cperey, ErikvdZee, Francois, jtandy, Leif, Linda, mburgoyne, MichaelGordon, Percivall, pvretano, Roope_Tervo, ssimmons
Regrets
Chair
Linda, Jeremy
Scribe
billrob__, billroberts, brinkwoman, Clemens Portele, ClemensPortele, Eric Boisvert, greenwood, Linda van den Brink

Meeting minutes

<brinkwoman> @tidoust just checking if you can hear us

<jtandy> francois - can you here us

<brinkwoman> \me rest are on irc

Spatial data on the Web BP implementation reports

jtandy: Work on SDW Best Practice completed in 2017 and helped to influenced several activities including WFS 3.0 in OGC. The SDWIG does not do standards, but helps to move relevant pieces of work forward.
… first item on the agenda are the best practice implementation reports to document implementations of the best practices

MichaelGordon: We now have 4 implementation reports (PSMA Australia, Kadaster NL, Geobasis NRW, OS). 70%-85% conformance.

ClemensPortele: The NRW implementation does not intend to reach 100% compliance as some BP aspects are not applicable.

MichaelGordon: Yes, this and that most of the implementations are existing implementations.
… BP document had only 1 minor correction in 2 years.
… More implementations potentially coming through OGC API development

https://‌github.com/‌w3c/‌sdw/‌blob/‌gh-pages/‌bp/‌BP-implementation-reports.md

<ChrisLittle_> A-

MichaelGordon: (shows implementation report pages on GitHub)
… Some of the "conformances" is really partial, but we can only express that in the text, the checkbox is yes/no

jtandy: Should we update the BP document to distinguish mandatory checks from optional ones?

Action: MichaelGordon to review best practices to suggest a list of the most important checks to the best practices and distinguish them from conditional checks due end of August 2019

MichaelGordon: (shows example "how to test" section in the best practice document)
… What do we want to do before close 2019?
… Or the end the end of the SDWIG...

brinkwoman: It is a good idea to add more recent examples. We also said a while ago that we use the reports to write an article. Updating the examples looks like it should have higher priority.

ClemensPortele: I will have a look at examples from the OGC API point of view.

billroberts: I will also have a review to look for new examples.

jtandy: What else do you have in mind?

MichaelGordon: I expect the OGC API activities could provide more implementation reports. It mainly is a task to identify ones and convince people to spend the effort.

billroberts: I am planning to create a report for the English Environment Agency.

MichaelGordon: Would the Weather on the Web work be another candidate?

mburgoyne: One challenge is that the dynamic nature of the data, which is a challenge for live links.

MichaelGordon: Yes, that would be a nice variation from the current reports.

jtandy: It would be a report on the UK MetOffice implementation (service hub).

jtandy: Roope_Tervo, what is the status of your implementation?

Roope_Tervo: It is in proof-of-concept state, it is unclear when it will be operational. I will have a look, if we could prepare a report.

MichaelGordon: If we achieve 5-6 reports from different backgrounds that would be good, so if we get two more we should be fine.

MichaelGordon: On the article, what would be the goal, just on the BP or the scope of the whole IG?

jtandy: I was thinking that we focus on the BP work. The article should be something that the press could reuse, a blog, etc.

MichaelGordon: I will create an early draft and send it out for comments.

Action: MichaelGordon to create a draft for an article on the best practices due end of August 2019

MichaelGordon: Open issues on GitHub...

https://‌github.com/‌w3c/‌sdw/‌projects/‌1

Hot topic: https://‌github.com/‌w3c/‌sdw/‌issues/‌1085

MichaelGordon: We agreed to see value and there are open issues how to address this

ClemensPortele: We have looked into sitemaps and it helps a bit, but the fundamental issues remain. There is also an event at JRC next week related to this.

Action: ClemensPortele to update #1085 with any conclusions after the workshop at JRC

billroberts: A general guidance how to deal with indexing of paginated resources with a large numbers of resources would be helpful.

jtandy: It will be up to communities whether it is valuable to index the individual items in collections.

MichaelGordon: (goes through list of open issues in https://‌github.com/‌w3c/‌sdw/‌projects/‌1)

jtandy: Any new input on #1086 (BP 14; describing positional accuracy)?

MichaelGordon: No new input so far

jtandy: Should we remove it then?

ChristinePerey: Microsoft has requirements for high accuracy in their HoloLens work. This might be a candidate?

(discussion that quality is generally important, and often information exists, but not in a web-friendly way)

MichaelGordon: It could be moved to the gaps in current practice section, if we don't find implementation reports.

Action: MichaelGordon to move BP 14 to the "gaps in current practice" section due end of December 2019

MichaelGordon: #1050 can probably be closed once we have the article

MichaelGordon: #1084 - Andrea proposed to discuss the DCAT revision during this meeting

MichaelGordon: #1037 is a bug that will be fixed in the editing

MichaelGordon: #1044 (conneg by profile) is on the agenda later

OGC API hackathon report

MichaelGordon: (reuses slides shown by Gobe in the OGC TC Opening plenary)
… OGC API Hackathon in London last Thursday/Friday, work on API implementations for Maps, Tiles, Coverages, Processes, Features
… OS is looking how people will interact with OS data in the future and sponsored the hackathon for this reason
… good spirit and good feedback from Geovation developers that are not geo-experts and are interested in the new specs
… also good that it appears as a common family of standards

ChrisLittle: It does not break anything done in features

ClemensPortele/pvretano: Yes, there are a few details that need to be worked out on the collection level where to draw the line between Common and the resource-specific specs

<brinkwoman> coffee break

hackathon (continued)

Mark: trying to make a lot of APIs consistent with each other might make it complicated and so chase away less expert users.
… You need to design an API with a target audience in mind

MichaelGordon: I had the opposite view and found the new proposed approach seemed to make it more understandable

Mark: coverages are likely to be used by specialists only. Note that the jargon can be confusing in itself, eg talk of Features and Coverages don't mean much to lay people

MichaelGordon: tiles and features might be the most commonly used aspects

jtandy: to summarise, we are balancing the desire of engineers to make things better vs what users can understand. So making sure we do user testing is done is important.

RobSmith: the current web map APIs are an interesting example. Eg mapbox, leaflet are easy for software people with little spatial knowledge, whereas the older versions of OpenLayers was harder for people in that community

MichaelGordon: UK government has set up a Geospatial Commission to look at geospatial strategy for the UK as well as running a series of projects. They are collaborating with various relevant public sector bodies
… There are 4 projects currently ongoing: Data Discoverability, Licensing, Linked Identifiers, Enhancing of the Core Data Asset
… the Linked Identifiers report has about 60 recommendations. It draws to some extend from the SDW Best Practices
… (the report is not yet available publicly but should come soon)
… Data Discoverability: user research was carried out with people with various levels of experience
… Main findings so far:
… - people use Google to find data (as opposed to government operated websites/catalogues)
… - CSV catalogues were produced and received a mixed response

jtandy: noted the CSV on the Web standards from W3C that could offer some useful facilities for this kind of thing

MichaelGordon (discoverability findings continued)
… some users want a clearer journey from catalogue to data

jtandy: there is a workshop next week at JRC covering this topic. Some of the outcomes of that will be linked back to this work

jtandy: can Michael persuade the Geospatial Commission to publish outcomes as OGC reports?

Michael: step one is getting agreement to publish openly, then we can think about which mechanisms

<jtandy> https://‌github.com/‌w3c/‌strategy

Strategy Funnel review

jtandy: we review things from W3C that have spatial aspects
… the W3C strategy funnel lists things that are being taken towards standardisation, showing which stage they are at
… here are the geospatial ones: https://‌github.com/‌w3c/‌strategy/‌projects/‌2?card_filter_query=label%3Ageospatial
… These include MapML, which is being taken forward by Peter Rushforth in a W3C community group and in OGC Testbed 15
… CityJSON: (a JSON version of CityGML v2) has reasonably broad adoption. Currently stalled and should perhaps be moved to the 'Parking Lot'
… the developers of CityJSON are not currently pushing for this to be taken through the standardisation process

tidoust: yes parking lot is fine if they are not pursuing standardisation

PROPOSAL: move CityJSON activity in the strategy funnel to parking lot

<MichaelGordon> +1

<RobSmith> +1

<brinkwoman> +1

<jtandy> +1

<Christian_Elfers> +1

<mburgoyne> +1

jtandy: no objections so resolved

Resolved: move CityJSON activity in the strategy funnel to parking lot

jtandy: next, Linked Building Data. Aim to harmonise with CityGML and other existing standards. This is mainly a communication activity
… WebVMT is on the agenda for tomorrow and Rob Smith will present in more detail
… SSN v2 - currently parked but may kick off again soon
… CoverageJSON is now incorporated in the OGC work on Web Coverage Service

<ted> [Linked Building Data Community Group is seeking to charter as a W3C Working Group, I'll be discussing with the chairs later this week or next]

jtandy: is anyone aware of other initiatives that should be added to this pipeline?

No additional items for the funnel were proposed.

jtandy: does anyone want to discuss OGC Technology Trends?

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

jtandy: note there is a 'Spatial Data on the Web' strand to the trends mind map and we'll cover those items during the meeting

Spatial Ontology/GeoSPARQL

brinkwoman: see http://‌ogcmeet.org/ There is a Geosemantics session in the OGC agenda for Thursday. It will cover a discussion of whether to create a new spatial ontology, or to further develop the GeoSPARQL ontology
… Frans Knibbe proposes have something equivalent to OWL Time.
… so come on Thursday if you want to contribute

josephabhayaratna (jo): a number of recent projects have shown some common problems and a group of people wanting to do something about it
… including Nicholas Car from CSIRO who has suggested some extensions to GeoSPARQL that can deal with some of the user needs. He will present on Thursday

billroberts: we use existing GeoSPARQL in practice

jo: can you send some more info on experience?

jtandy: please send via public mailing list

CRS Content Negotiation

jo: (sharing screen) various working groups in OGC and W3C have been looking at related items, eg the Data Exchange Working Group
… There are issues with geojson for example, which currently has to be WGS84, but that's not always the best choice of CRS
… It has been suggested that conneg could be used for CRS in a similar way to DXWG ideas on conneg for application profiles
… Need something like a register of CRS that could be used in implementations. An HTTP Accept header would be needed
… with unambiguous ways to refer to the CRS
… Jo has been looking for people in OGC who could help with that

(we can hear you again now - there is a delay on the line I think and some intermittent connection)

(lost you again)
… note lots of contributions. Jo will share draft

jtandy: on behalf of SDWIG, we'd be happy to help share this info - eg via public mailing list.

jo: will ask contributors first just to check attribution is properly included, then put it on the public mailing list

jtandy: suggest we put this in the Exploration phase of the strategy funnel

jtandy: Francois, please work with Joseph to get this into the Strategy Funnel

jtandy: anyone here want to help progress ?

Eric Boisvert has volunteered to help

and Rob Smith thinks Peter Rushforth might be interested to help so we should ask him

<brinkwoman> meeting ajourned until 4:30

<brinkwoman|2> can you hear us?

<jtandy> hi... we are just trying to resolve audio issues

future direction of sdwig

<billroberts> https://‌www.w3.org/‌2017/‌sdwig/

<jtandy> jtandy: at w3c groups are chartered for a limited time period, in this case until end of 2019

jtandy: this group end 2019. Look at charter.

jtandy: need to decide what to do. OGCTC good occasion to discuss and decide. typicall 2 year cycle

jtandy: to fix in w3c cycle. 3 years cycles. This expected to be first discussion

jtandy: lots of dividen from association OGC W3C on geospatial issues. biggest for OGC : development of WFS3

jtandy: wants to propose to keep going.

ssimmons: OGC don't time limit. no worry about same constraints. hate to see any hurdle. very useful for us

ssimmons: impacting a lot of things. O&M, WFS, lots of influence. changes overall authority. ISO not the only authority

boyan: raised visibility of spatial

jtandy: we all agree . what basis does that take ?. from chair view. best practice stuff, few people . 1 task 1 person

jtandy: often 1 person, 1 task. Difficult to create a coherent group. Hard to get time to meet because time zone

jtandy: moved to "focus week". Good turn around on issues.

Linda bringing best practice to screen

josephabhayaratna: question "inaudible"

jtandy: hot topic identifies. last 3 months - almost no work. not well functionning from a chair. Looks like individual moving issues

ted: background. is it because of availability, interest . works in isolation. Better narrower scope ?

jtandy: reduce scope and do fewer things ? As a group we lack coherence similar to best practice. Had impact on everybody

jtandy: less distraction ?

jtandy: we created a mechanism to have diversity

<ted> [if there is a prioritized queue of topics, then those with topics further down the list will have some motivation to help with the ones ahead of it]

<ted> [even if not their core interest/expertise]

ssimmons: you are correct. best practice was the only product. standard were different. lots of focus. Large group . more exciting to work on BP

ssimmons: lots of standard starts with 1 person. not unusual

<ted> [alternately how do we find additional people with interest and expertise in specific topics to draw in on furthering the work?]

billroberts: agree jt. Kind of darwinian only one left of then group.

billroberts: Not of a lot of people on the group who want to work on 1 person thinks ?. hard to know. start small and grow. various reasons.

billroberts: maybe having focus where significant contribution

mburgoyne: Once you start going to standard. Taking from opposite view. Is is worth include in OGC

<ted> [criteria for consideration of a topic being taken up could be critical mass of interested parties > 1]

Christine Perey: example of other example outside w3c is immersive

jtandy: people who wants to bring AR in browser

jtandy: same for spatial ancher. apply ancher in real world in video

jtandy: another one (calling ted) : autonomous vehicle. looking into in-car stuff. prediction: outside will be important

jtandy: another example: future vision of this group. rather trying to help people, maybe we should coordinate. You know what Immer Reality group. We have gaming group

jtandy: role of this relationship to coodinate and overlap share interest

ssimmons: coordination is really import. we got cue on autonomous vehicle. waiting for standard. OGC now attending

ssimmons: coordinate will become important. we should be contact

ted: talking about autonomous vehicle charter. concentrate on data available to vehicle *cut*

ted: roolback : have had strong interest LBS for vehicle . algees ? we have now useful information, we can create ontologies. We are in the peripheries, we can provide data

Christian_Elfers: what is coordination effort. Active role in attending. interjecting, introducing in those groups

jtandy: I see a coordination group, and I see that other group interest. Like a dating agency. You both looks like you would go along great

jtandy: cross polinisation and do the introduction. Where we have seen success is where we have seen interesting body of work into another

boyan: struggling. value was really in the product in the working group. If the problem is to generate more participant. not sure if you'll get more participation with coordination

jtandy: wg was charted to deliver particular product. interest group is to incubate to get ready for "chartership"

jtandy: what we are trying to do. In order to get to standard, need support to get the work done. main difference between two

jtandy: for encourage participation. if the group plan to do work, you need more participant. But it might be a different level of work

boyan: smaller body coordinate, larger body does the work.

jtandy: In order to drive participation, the reward is tangible thing. Best practice or standard

boyan: no commitment - don't do it

boyan: group might lull until they realise they need to do something

jtandy: your point is inactivity until we agree on issue

boyan: still need coordination to lead

jtandy: if we realise a group of people interested - might tell them to create a one

mburgoyne: Might not aware of existance

ssimmons: two great success. understanding operation and business to generate the right amount of excitement

ssimmons: we have to find work to generate sponsor, align to their business. ISO was opposite. OGC had to do all the work to hand them over

ssimmons: We have to probably one more round to understand excitement generator

ted: first. pro and cons for each. criteria, 5 supporters, then github. sort of incubator group. interest group might interest to follow that model. interest group try to generate excitement

greenwoood: sorry for bad notes

<ted> [maybe IG can further the Linked Building Data Community Group (CG) model. to form a CG at W3C someone proposes an idea for potential standardization, at least N (5?) need to endorse it]

jtandy: what interest group has provide is safe place to spin idea in public. if you want to share, where do you start. start github in NRCan?. This liaison example is a github place

jtandy: example. CRS content negotiation. No one looking at this. So this group provide a space to test idea and look for support and consensus

jtandy: my perspective. in addition to coordination, place to test

<ted> [the IG can help propose those ideas, align with OGC and help promote them. review which are ripe for going formal standards route]

jtandy: if nothing happen , chair will find people to work on

Christian_Elfers: appologie for not attending for some year. as outside, see the work,specially BP, taking care of link technoloy and geospatial

Christian_Elfers: great if you could focus on those topic to be sure they fit. Lots of value from document

Christian_Elfers: Has to be concrete. has to be useful, covers both worlds. no example in mind, but there must be thousands

jtandy: This is about identifying a product

ssimmons: one of the way we generate idea is innovation program, become IE, pilot. Idea are put in git, everyone can comment. sponsor look at issues and propose pilots

ssimmons: maybe this group could consider issues. AR is sitting in there waitinh

jtandy: More of coordination.

ssimmons: explain how to issue are significant to generate more activity

boyan: more down the road investment

ssimmons: majority are funded because someone has an idea and someone wants to fund it.

boyan: still time need to evaluate issues

josephabhayaratna: lots of issues from pilot and potentially look through lenses of wfs. More a case of review of documentation and responding to issues

ssimmons: good job Joe

josephabhayaratna: hard to follow audio (greenwood: i share your pain)

ssimmons: on the tracker, involve people to comment , throught the lense of wfs. 78 open issues

jtandy: ask to put url in minutes

<ssimmons> OGC Innovation Program ideas issue tracker: https://‌github.com/‌opengeospatial/‌ideas/‌issues

jtandy: ask we migth go about activities. 3 types of activities. 1: coordinating between OGC W3C. 2: provide space to share early ideas. 3: proactively review issues on tracker

jtandy: from W3C perspective: strategy funnel is equivalent

jtandy: what kind of group needs to do that. OGC staff activity ? ssimmons agrees

jtandy: push to OAB . cross view of all sort of things. they are informing the TC. Good place, but not much resource

ssimmons: quarterly SDW discussion in the OAB

ted: can be the conduit on W3C side

jtandy: OAB, here the list of things we consider

jtandy: On W3C : tag group. they meet monthly.

ted: unsure, either monthly or bi-weekly by phone and 3-4 f2f a year. 3 or 4 F2F

jtandy: by teleconf

jtandy: fold in Geosemantic group ? thing could be spun up in GS group (same people in both group)

jtandy: fourth option: entirely new group

boyan: 2 model. coordinating and model

jtandy: if we want to build product, we need a group to do that. SWG, DWG or join, I am not seeing a significant group to contribute

boyan: in term of ref. you don't want to exclude

jtandy: chart for 2 year. If you spot something, chart it

ssimmons: that what we would do

jtandy: how much support you would give to someone with new idea. But you right, need charter, etc.. but no product identified yet

linda: unless geosparql update, it's OGC , logical to remain OGC, but not required

boyan: exposure of W3C brings a lot

billroberts: as long a skill in OGC ok, but lots of ontology in W3C

boyan: some work in W3C is not happening here

jtandy: you'd need to create a charter and a that point anyway

linda: interest group, not doing standard

jtandy: Can go join group again and benefit

jtandy: IF there is support in both camps, or just one, to put time and product, it will fly. But we don't have so far. Need to create a new SWG

boyan: confused. 2 or 3 activities part of this group, product oriented. GeoSparl, Sensors, ..

boyan: matchmaking and work will be done anyway ?

<ted> [incidentally we are using SSN in the vehicle signals ontology VSSo]

jtandy: geoparql update : just incubating and pushed on geosemantic as best place. SSN , we puiblished a note, extension, no standard

ssimmons: OGC discussion paper like

boyan: if we are just coordinating, need to push elsewhere

jtandy: agree. it's been pulling teeth.

boyan: product great, need to be done. if you can create some and not other, slippery

jtandy: recharter in a way that just coordinate and push to other

jtandy: CRS also incubated - looking for a group

jtandy: video tracking, Rob on his own

boyan: misunderstood you were just match making

jtandy: 2 years ago, hope we created more product. did not. need to review chart

josephabhayaratna: wondering mention long ago to prioritise on small number of thing good sandbox things that would fly

jtandy: exception of BP, we haven't anything with more that 1 person interest. very short list.

ssimmons: not lack of trying

josephabhayaratna: is it just people running their own list intead of having a list to pick from collectively. Too much work in progress ?

jtandy: my feeling. not enought benefit for host organisation to carry work.

jtandy: nobody thinks other project is important for own priority. looking for thing contribute to their organisation

billroberts: agree. day job, commit to something that match priority. don't want to do too many things

jtandy: autonomous vehicle broad enough to get fruitfull product. but not on our book

boyan: would not do it anyway on new charter

jtandy: jo agree ?

josephabhayaratna: not sure

boyan: agree with jtandy

billroberts: we do networking. if you want to know, you know who to talk to. Not sure coordination pipeline will do better. What is valuable in working with this group. learn a lot more about OGC. learn about what and who

billroberts: network not enough to keep the group, but necessar

ted: (question did not get)

<josephabhayaratna> josephabhayaratna: is it really the second area: having a safe place to share ideas early

jtandy: intrinsic value in coordination is seeing what people are doing by offering time to read other people stuff. is it enough ?

jtandy: I think that we had a good discussion to get idea. reasonnable consensus about the value of the function and the need to separe coordination from product generation

jtandy: we as a group, it's probably first of many conversation. Action on Linda and jtandy and go read back in the minute and synthesise proposition and spell this is what we mean

boyan: emphasise coordination model potentially will reduce participation because some want to focus on coordination, other on products

jtandy: Good point. needs to go on annouce. One further thing to ask. need to go W3C and OGC staff for question

ted: transportation data

<Zakim> tidoust, you wanted to reformulate review option

françois: review the options: what I hear. similar group to TAG for geospatial. they meet often. 5 or 10 years ago, was not that well. Place people can come for review on technology

jtandy: not totally what in mind. need to synthesise.

jtandy: charter says we have one meeting, next in Japan. Are people planning to get to Japan.

ssimmons: Someone from OGC will (Ingo ?)

jtandy: for coordination

billroberts: unlikely

jtandy: Linda and jtandy are rep on W3C, may go anyway. But if noone else, moot to have a meeting, teleconf can be anytime

jtandy: Maybe fall meeting in Toulouse

boyan: is it needed

jtandy: great valid to do

billroberts: Great idea come from meeting

jtandy: if not sufficient attendance, no meeting planned.

jtandy: thank you for attendance

jtandy: resume at 13h00 tomorrow

meeting closed

<brinkwoman> Thanks everyone, until tomorrow!

Summary of action items

  1. MichaelGordon to review best practices to suggest a list of the most important checks to the best practices and distinguish them from conditional checks due end of August 2019
  2. MichaelGordon to create a draft for an article on the best practices due end of August 2019
  3. ClemensPortele to update #1085 with any conclusions after the workshop at JRC
  4. MichaelGordon to move BP 14 to the "gaps in current practice" section due end of December 2019

Summary of resolutions

  1. move CityJSON activity in the strategy funnel to parking lot
Minutes manually created (not a transcript), formatted by Bert Bos's scribe.perl version Mon Apr 15 13:11:59 2019 UTC, a reimplementation of David Booth's scribe.perl. See history.

Diagnostics

Succeeded: s/Christian_Elfers/Christine Perey

Succeeded: s/AI/AR

Succeeded: s/things/thinks

Succeeded: s/meet bi-weekly/unsure, either monthly or bi-weekly by phone and 3-4 f2f a year/

Succeeded: s/ted/TAG

Succeeded: i/Mark: trying to make a lot/scribe: billrob__

Succeeded: i/Mark: trying to make a lot of APIs/scribenick: billrob__

Maybe present: boyan, ChristinePerey, ClemensPortele/pvretano, françois, greenwoood, jo, josephabhayaratna, Mark, Michael, PROPOSAL, RobSmith, ted, tidoust