24 Jan 2018



Kaz_Ashimura, Zoltan_Kis, GraemeColeman, Darko_Anicic, Michael_Lagally, Kunihiko_Toumura, DarkoAnicic, Daniel_Peintner, Toru_Kawaguchi, Michael_McCool, Michael_Koster, Matthias_Kovatsch, Sebastian_Kaebisch, Soumya_Kanti_Datta, Tomoaki_Mizushima, soumya, dsr, Ben_Francis, Dave_Raggett, Yongjing_Zhang, Ryuichi_Matsukura
Matthias, McCool, Yongjing


<mkovatsc> https://www.w3.org/WoT/IG/wiki/Main_WoT_WebConf#Agenda

<kaz> scribenick: GraemeColeman

Quick updates

mkovatsc: Dave was to present something on the agenda, but he is not yet on the call
... Any updates?

Soumya: Has circulated a call for papers - if anyone wants to write a white paper describing the current status, that would be highly welcome.

mkovatsc: We need someone in the group to sketch some concrete ideas as to what a submission could look like - paper, book, demo. This should go on the agenda.

<mkovatsc> IoT Week, Bilbao, 4-7 June 2018

mccool: Thinking about writing up some of the work on security.

<kaz> Call for Papers - workshop on semantic interoperability

mccool: Suggest talking offline about writing one large paper other than many smaller ones.

mjkoster: Asked if a submission can be submitted if published somewhere else?

Soumya: No, that is not possible.

mccool: Still possible to publish in the journal later even if published in a previous venue.

Soumya: To mjkoster - you can expand the conference paper and submit it to a journal. This is standard for IEEE.

mjkoster: Putting together a paper on design patterns for interoperability, which will cover WoT as well as general web media. Currently thinking about whether to write this up as a single paper or multiple separate papers.

mccool: Worth throwing around a bunch of ideas and seeing what works together well. If there's more than one paper, that's fine, but if they fit together well, then a single paper would do. But IP protection issues, and timeline, also need to be considered.
... Let's follow this up on the Thing Description TF meeting, as that's where this discussion is relevant.

<dsr> Here link to iotweek http://iotweek.org/

<kaz> Soumya's message on the workshop

dsr: I was asked if we were interested in doing a workshop (0.5/1 day) on IoT on the WoT on challenges and opportunities within this field and in the web of data. Would involve panel sessions, demos, etc. Audience would be from a broad range, but would include people from European research projects.
... Has been involved last week.
... last year.
... It would be a good opportunity to make a clear description of what we've achieved in this group, to disseminate our results, and to perhaps receive more support from people in the audience.
... The topic of marketplaces keeps coming up in EU research projects - so how do you enable open markets in IoT? We could cover this in our workshop, as to how this fits into the web of things.

mkovatsc: Other projects involving marketplaces are going, so (as there are many academic partners) they may also be interested in getting involved in the workshop.
... This should be a sticky for the main call (as to whether main contacts would be interested in getting involved).

Tooling and practices for W3C web data standardisation

<dsr> https://www.w3.org/2013/data/ and https://www.w3.org/2017/12/odi-study/

dsr: Has taken over as the activity lead for the Web of Data. Has conducted a survey of tooling and practices funded by the UK government.
... Created a lengthy questionnaire to get a broad range of feedback from a wide range of responders. Received around 75 responses. Questionnaire took about an hour to fill in.
... Wanted to understand what people from different backgrounds were interested in - e.g. sustainability and governance (e.g. best practices on how to proceed), scaling (translating overlapping categories and what the practices are), tools and practices (what can be used to support those practices), links between communities.
... "How successful are standards"? What is the level of interest, and what is the adoption level? Logs taken from the W3C website - there is still considerable interest in JSON-LD.
... A lot of interest from US and China.
... How are we going to measure what the WoT is?
... Please look at the report, and we can discuss in a future meeting when people have had a chance to review it.
... Companies are looking at IoT platforms and platforms-as-a-service, but not necessarily looking at open markets for services. Yet, based on Google/Apple app stores, there is potential for linked standards for marketing.
... What changes do we need to make at the W3C? How could they fund and support standardisation? This needs to be discussed with stakeholders and W3C members. Does it make sense to change the funding model for community-based standards?

mkovatsc: Are you planning to provide publicly available metrics for the W3C? Or do we need to develop tools ourselves?

dsr: A bit of both. If we can get the support from the members, we could look at interest over time in a particular standard. We need support of the systems team for that, but they are fairly stretched. Beyond that, how do we gather the information? How many people are implementing the standards? Can we find ways of helping people registering their interest in using those standards?

mkovatsc: Did you identify the most popular specifications for open data - e.g. most popular formats for providing/describing open data?

dsr: Don't have a specific answer - but there is a lot of interest in comma-separated values, whereas others are using e.g. JSON-LD. These results pose the question how do we deal with this?

mkovatsc: So there is no specific information as to which is the most popular?

dsr: We did ask which standards were being used, and whether they were interested more in stability or agility, and the problems around interoperability/different licenses. Second phase will dig deeper into these results.

mkovatsc: It would be interesting to find out the most popular formats, and how they do the description so that we can compare it to what we do.

dsr: Next stage will be to find out the priorities for a second phase, and to secure funding to do this.

face-to-face meeting in Prague

mkovatsc: Offer from Sheraton Hotel to do the meeting. 85 Euros per person per day. Assuming an average of 40 people, cost will be around 20,000 Euros. Combined with not having a specific host. Looking for sponsor - internally, this was felt to be a bit expensive, so possibility of looking for a second sponsor, or setting up participant fees. Or, just find a different host.
... Another approach is to look out for other hotels - a smaller hotel might be possible because we don't need big meeting rooms.
... Please ask around if you would be willing to host/co-host this event. Costs expect to be 20,000 Euros, but we hope to reduce this depending on whether or not we can find a co-host. We want to avoid registration hassles.

mccool: What days are being used to calculate these costs?

mkovatsc: Saturday to Thursday. Average of 40 people, as there will be fewer people at the Plugfest. But, it highly depends on the location and how many people will show up. It's a rough estimate.
... The day rate is per person, per day. Includes the meeting room in the service costs per person. We can go back to them and ask about a fixed price for the numbers who actually show up.

mccool: I'm assuming the Friday meeting will be paid for by OCF - but we will need to figure this out.

mkovatsc: I was expecting that we would join the co-meeting with OCF, so it wouldn't be on us, but we would use their meeting rooms.

mccool: We would need to check this.

mkovatsc: Any other opinions on what we should do? Should we look for another host - or end up with a meeting fee for participants?

mlagally: Maybe we should have a Doodle poll to find out how many people can come?

kaz: I will create a questionnaire to find out how many people will go.

<kaz> ACTION: kaz to create a WBS questionnaire for Prague f2f

<trackbot> Created ACTION-126 - Create a wbs questionnaire for prague f2f [on Kazuyuki Ashimura - due 2018-01-31].

<dsr> http://www.meetings-conventions.com/Meeting-Facilities/Prague/Hotels

dsr: Potential for self-hosting in Prague - look for cheaper, "tourist" hotels with meeting rooms.
... There are some websites and places (such as the above) which may help find some initial candidate hotels.
... Firstly we need to find out people will attend.

kaz: Let's have more discussion about this on the Chairs list.

Task force reports

mkovatsc: When do they plan to release the next working draft?

<kaz> [TD]

Sebastian: Thing description TF - we had some discussion about simplified things description proposal submitted before Christmas. We discussed about interaction patterns, property/interaction events, are there enough. Matthias and myself held a separate call with Carsten Bormann about his concerns regarding property/action events.
... We discussed what a property means, what an interaction means and so on - we need a more concrete definition, and examples, for each. This will help overcome Carsten's concerns.
... Plan for next meeting - we will concentrate on mkoster's proposal on new vocabularies. Also giving a status report on the TD.

mkovatsc: It appears that the interaction model is not commonly understood in the same way by everyone. Possibly because properties, actions and events are quite fuzzy, as they are similar terms to elsewhere. People therefore need guidance for modeling new system. So it's not specifically giving definitions, but about giving a description about how the interaction works and the data action model works, what the idea behind it is and so on.
... With the simplified TD, the different ideas are not well documented, or are included in a big PR thread which has now been merged, so very difficult to follow.

BenFrancis: By having the defaults, is it possible that all of the semantic markup (JSON-LD tags, etc.) could be optional? e.g. similar to OpenGraph data which is added to plain HTML. You have HTML and then you add the extra data. So, could you have a plain JSON description, and then the semantic information is optional.

mkovatsc: Biggest problem is that LD doesn't support defaults or optional information.
... Whatever is encoded in the TD must be machine understandable and readable. So, we decided to take an existing spec (JSON-LD), but this obviously has some limitations. Let's discuss this in the TD call.

DarkoAnicic: Question about defaults - even if we find away to dismiss them in the TD, they still need to be implemented. This means that every client needs to implement these rules to be aware of the defaults. So I'm not sure if omitting them would be of benefit, because clients still need to be aware of the default values.

BenFrancis: These defaults can be separated out. Problem is that if it's too complex for humans to read, we have a problem of adoption. Doesn't matter how good machines can read it.

mkovatsc: That said, it's quite easy to hard code defaults in some piece of software.

BenFrancis: Look at web app descriptions. These are quite widely adopted and understood.

<kaz> Web App Manifest

<kaz> [Scripting]

zkis: We discussed agreed Scripting API to be used at the Plugfest.

<kaz> [Binding]

(Koster had to leave due to a meeting conflict. also we were out of time)


Summary of Action Items

[NEW] ACTION: kaz to create a WBS questionnaire for Prague f2f

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2018/01/24 16:49:25 $