W3C

- DRAFT -

WoT-IG/WG

04 Dec 2019

Attendees

Present
Dave_Raggett, Michael_Lagally, Kaz_Ashimura, Michael_McCool, Christian_Block(eCl@ss), Ege_Korkan, Kunihiko_Toumura, Markus_Reigl(Siemens), Taki_Kamiya, Daniel_Peintner, Sebastian_Kaebisch, Tomoaki_Mizushima, David_Ezell, Michael_Koster, Ryuichi_Matsukura, Takahisa_Suzuki, Tetsushi_Matsuda
Regrets
Chair
Sebastian
Scribe
mlagally

Contents


<kaz> scribenick: mlagally

Agenda

Ege: additional agenda point: Marketing TF resolution to operate independently

Sebastian: independent from WG?

Ege: independently from main call to reach decisions

McCool: additional point: logistics of call timing

Kaz: this week and the next there will be important Project Reviews (one on the Repository Manager, another on the W3C Process). So cannot join the Marketing calls.

Quick updates

<ege> Happy Birthday Sebastian! :D

Sebastian: Markus Reigl (Siemens) and Christian Block (eclass) are on the call
... they work on alignment with the TD
... <walks through agenda>

<kaz> Agenda

Date/time for the joint call with MEIG

<kaz> doodle results

Kaz: based on the doodle results above, the joint discussion with the Media&Entertainment IG will be held on 4 February 2020 during the MEIG call
... I'll send a reminder with the concrete WebEx coordinate to you all.

Joint call with eClass

Markus: most of us met before, I'm heading the Siemens standards and regulation business
... Siemens is involved in other SDOs
... we are also member of eClass EV.
... This is a non-profit organisation to define an XML based standard to describe assets
... The standard exists already since 20 years
... primarily used for B2B and B2C, logistic applications, supply chain on purchase side to classify and describe a product in electronic catalogs
... representing a taxonomy based, tree based system for provisioning of dictionaries in more than 30 industrial domains, e.g. medical, production, food and beverage
... frequently used to structure catalogue data to be widely used
... exchanged between value / trade partners
... The classification part has 4 layers
... <shows an example of a web shop for fashion items>
... <another example: screws, bolts and nuts, which may have additional properties, e.g. sizes, head shape, ...>
... another example: a motor that has different classification data. This data can be retieved from a vendor-provided database that follows the eClass standard.
... we call that almost a "digital twin", it has hundreds of properties
... another example: DC power supply, that is classified using the 4 layers. In addition it contains the engineering data.
... Let me explain the eClass orginsation
... Primarily an european outreach, used in France, Netherlands, Spain, ...
... currently only little use in Japan and PR China
... Head office has around 6 people, there's a CRM and CRD for management and design of requirements for the next release
... releases typically once a year
... there are 45 expert groups, e.g. medical, electric goods, electric controls, sensors
... 140 member companies, 3000 applications, even in the US we observe some downloads

Lagally: what does an application mean?

Markus: usage of the specification by a company
... properties describe all relevant aspects of a device, this goes down to almost a digital twin
... segments - main group - group - subgroup
... under the subgroup you see the properties
... eClass is very advanced in terms of classes and properties - German and France have contributed heavily
... define semantcs for Industry 4.0 use cases
... This is why eClass has approached the W3C
... eClass standard can rovide a very rich classification and property repository that can be used for thing descriptions

McCool: we have to see how to use eClass vocabulary

Markus: Christian has been working with other W3C members

Christian: LEt me summarize technical aspects of eClass
... >40000 classes are managed in a online catalog
... we intend to bring this catalog with rich semantics to the thing description
... application classes and proerties can have synonyms as keywords
... advanced concepts: we have a bundling of properties
... e.g. bundle all properties of one manufacturer
... we offer object oriented concepts such as cardinality and polymorphism
... all is encoded in XML and linked
... at the moment it is a rich dictionary which describes also semantics
... we had a kick-off meeting in Munich in June
... since June we had 8 meetings
... we analyzed how to incorporate eClass in thing description
... JSON-LD linkage to eClass dictionary
... as a step further we may need an RDF/OWL representation
... data model is based on an ISO norm, 3 companies have tried to do a model transformation
... transformation of advanced concepts causes some issues
... we are working on transformation to RDF/OWL
... we plan for a F2F for 2020

Kaz: this work includes 3 pieces, not only TD but also data schema and data catalog, and I think we should think about all those 3 pieces.
... in addition, recently W3C started decentralized identifier work, and that may also be applicable
... further collaboration would be useful

<inserted> (Koster mentions "very interesting and would like to follow up on intersections with iotschema and OneDM definitions. Unfortunately I need to drop ff now for another meeting".)

Lagally: did you find any gaps / new requirements for TD or WoT architecture

Sebastian: I'm positive we can address the use cases

Lagally: how do you describe semantics / digital twin?

Markus: digital twin has different interpretations, you may need several hundred properties, eClass defines a structured way to describe properties

Lagally: are you also targeting other geos?

Markus: in manufacturers in China and Japan, Alibaba, Granger and Amazon in the US

Lagally: thanks!

Ege: linking of a TD to an eClass would be very beneficial, is this already possible?

Christian: references to the dictionary via an IRDI is already possible

Ege: in a deployment scenario, I may be able to generate a TD automatically
... are there any plans to host eClass descriptions on the device?

Christian: the device can provide the class identifiers and then can be looked up

McCool: we are running out of time, continue in the next TD call?

Sebastian: we may have a separate group that writes a guideline document

McCool: suggest to start in the TD call
... we need to talk about PR resolution

<Christian-Block> the current eCl@ss metamodel as mentioned: http://wiki.eclass.eu/wiki/ISO_13584-32_ontoML

WG Charter

Sebastian: deadline is in two weeks - we have 11 supporters, Adobe suggested some small changes
... let's revisit in next week's call
... if your company has not yet voted, please ping your AC rep

PR resolution

(Lagally has a problem weith WebEx and Kaz takes over the scribe)

scribenick: kaz

McCool: agree that it would be confusing to have both readOnly and writeOnly

Lagally: as Ege pointed out, we're inconsistent with TD and Architecture

Sebastian: where the specs mention them?

Ege: the property is subset of data schema

<kaz> 5.3.2.1 DataSchema from TD spec

McCool: we can't change the spec given it's already a CR
... can we live with the current spec?

(no objections)

McCool: the CR2 period for TD is over now
... and the question is if we're ready for a resolution for PR transition?

McCool: should we have the resolution today, otherwise we may run into timing issues

scribenick: mlagally

Taki: in the TD spec there's still CoAP content part defined as t.b.d. - we should fix this before going to PR

scribenick: kaz

Kaz: we need to clarify that all the Issues and PRs are resolved (or marked as "v.next"). also we need to display all the features are implementable using the Implementation Report.

scribenick: mlagally

Sebastian: we could have the PR resolution on Friday

McCool: we may need to do updates to remove the at risk items

Kaz: based on the updates, we need to talk with Ralph and PLH prior to the transition.

Future F2F

<McCool> mccool: I will have to drop at half-past, regrets in advance

Sebastian: Mozilla offers to host at different locations
... We should go for an option that allows for many people
... Mountain View can host most people, let's decide next week
... June/July potentially hosted by Intel in Helsinki

Marketing TF reports

<McCool> mccool: sorry, i have to drop; re marketing resolution, I vote "yes"

<discussion about whether marketing is empowered to make resolutions about communication to the outside world by themselves

Lagally: prefer to present significant material to the main group and ask for approval. Minor changes, blog posts, Wikipedia etc. are fine.

Taki: I suggest to post to the mailing list, so everybody can review beforehand

<sebastian> proposal: are there significant changes that should be shared to the public (e.g., white papers, slides) there shall be a vote during the main call (a notification to the group should be shared beforehand), minor changes like blog posts, wikipedia etc can be changed directly by the marketing TF team

RESOLUTION: are there significant changes that should be shared to the public (e.g., white papers, slides) there shall be a vote during the main call (a notification to the group should be shared beforehand), minor changes like blog posts, wikipedia etc can be changed directly by the marketing TF team

WG Charter (revisited)

<Zakim> kaz, you wanted to mention PR910/910 for the wg charter draft

Kaz: Sorry but we didn't explicitly confirm the recent changes to the WG Charter draft, so please let me rivisit that quickly.
... I have created PRs to update the WG Charter draft using the text from the new Charter template based on our agreement during the previous call

PR910 and PR911 for the update

Kaz: got approval from assigned reviewers, so will merge them

Marketing call slot?

scribenick: kaz

Kaz: regarding the proposal to change the date/time for the marketing call to avoid the overlaps with the Project Reviews, I'll send a message to the group list

Architecture call

Lagally: there will be an Architecture call tomorrow on Dec. 5

[adjourned]

Summary of Action Items

Summary of Resolutions

  1. are there significant changes that should be shared to the public (e.g., white papers, slides) there shall be a vote during the main call (a notification to the group should be shared beforehand), minor changes like blog posts, wikipedia etc can be changed directly by the marketing TF team
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/12/05 14:38:17 $