Web of Things IG

31 Aug 2016


See also: IRC log


Kaz_Ashimura, Dave_Raggett, Harada_Kei, Andreas, User2, User4, Danh_le_Phuoc, Harada_Kei(again), Klaus, Martin_Bauer, Nicolas_Seydoux, Peter_Kim, Soumya_Kanti_Datta, Takuki_Kamiya, Tim_Carey, Victor_Charpenay, Jaeho_Lee, Omar_Elloumi, SeungMyeong_Jeong, Thomas_Kessler, Masato_Ohura, Antonella, Ian_Blair, Katsuyoshi_Naka, Kazuaki_Nimura, Mengxuan_Zhao, Wonsuk_Lee, Xueqin, Yongjing_Zhang, Catalina_Mladin, Chonggang_Wang, Hiroyuki_Maeomichi, Carmelo_Zaccone


<mkovatsc> https://www.w3.org/WoT/IG/wiki/IG_WebConf#Agenda_of_next_WoT_IG_WebConf:_31_August_2016

<MartinB> Hi

<scribe> scribenick: dsr

<kaz> agenda: https://www.w3.org/WoT/IG/wiki/IG_WebConf#Agenda_of_next_WoT_IG_WebConf:_31_August_2016

WoT: Introduction of the TD

(no link to slides)

<LindsayFrost> If the slides are uploaded, please mention the link here ...

<mkovatsc> thx Kaz

Darko presents a short presentation on thing descriptions. (Link to be added shortly)

Things as having a resource model, protocol bindings and server connector

The resource model describes the interface exposed to apps

<McCool> privacy?

Darko shows a diagram where a client is connected to a server for remote access to a thing

<McCool> re privacy: noted

The client retrieves the thing description to discover how to expose the thing to a local app, and how to connect to the remote thing exposed by the server

Darko presents a diagram depicting different categories of metadata in a thing description (resources, semantic models, communication and security)

<mkovatsc> https://www.w3.org/WoT/IG/wiki/OneM2MCall31Aug2016

We’ve been using JSON-LD for representing thing descriptions. This offers familiar JSON syntax plus rooting of terms in RDF

RDF allows us to exploit semantic tool chains and ontologies

Victor presents an example thing description for a thermostat

<McCool> where does timing data get represented? For instance, devices that sleep; long-running actions; sync vs async actions

<McCool> will follow up later, no need to interrupt pres

This exposes the current temperature as a read-only property, raises events when the temperature rises above a given limit, and an action to to set that limit

<McCool> ok

Victor explains that this example provides a binding to the coap protocol

<kaz> [ fyi, Tech Landscape for Security & Privacy: https://w3c.github.io/wot/landscape.html#tech-landscape-for-security-privacy ]

<Zakim> McCool, you wanted to ask where timing data gets represented. For instance, devices that sleep; long-running actions; sync vs async actions

Michael McCool asks about where timing data gets represented. For instance, devices that sleep; long-running actions; sync vs async actions?


Johannes adds to Victor’s explanation of long running actions

<mkovatsc> (Johannes)

Darko: we currently don’t time stamp events, but this could be added or passed explicitly as part of the data model for an event

Explanation of the use of a repository for thing descriptions, along with protocols for registration and discovery

We have experimented with using SPARQL for querying a repository

Explanation of use of rich semantic models, e.g. based on ontologies sich as SSN, QUDT and SAREF

Victor shows relationships between thing description and SAREF

He shows a script that executes a SPARQL query to the respository


Yongjing: thanks for arranging this call and for the presentation

I have a question on the semantic extension part

<kaz> (Extending Thing Description with Other Semantic Models)

Do you have a specification for how to use thing descriptions with other ontologies?

Or are you reexposing external semantics in a thing description?

<kaz> [ fyi, Current Practice document: http://w3c.github.io/wot/current-practices/wot-practices.html ]

Darko our spec is the current practices document.

This describes what we do for each plugfest we hold

Michael Koster talks about the details for how the annotations are represented

<kaz> ack q?

Matthias: we’ve shown how to import external vocabularies

Darko: this uses the @context feature from JSON-LD

Yongjing: nice diagram to show distinction between domain dependent and domain independent models

Tim_Carey: it will be interesting to see how we (oneM2M and WoT IG) can interwork

oneM2M: Semantics presentation

OneM2M started with a domain independent approach for services, but our interoperability requirements point to need for domain dependent models too

HAIM is acronym for home appliance information model

Yongjing presents a UML model diagram

This includes static metadata, module class version numbers and so forth.

We have a separate UML diagram for data types

We handle units of measure and constraints as part of the data type model

HAIM release 2 covers 14 devices with 41 module classes. Television as an example

TV resource tree with leaf nodes including serialNum, deviceType, audioVolume etc.

We’re looking at mapping to models from other IoT organisations, e.g. OCF and Alljoyn

We expecting to discuss extension to other domains, e.g. vehicular

For collaboration with W3C, we could look at alignment of models (HAIM & TD), as well as using HAIM to validate W3C WoT TD against a home appliance model


Daniel: what is your process for new module classes?

The TV example includes properties, actions and events

Johannes: You have devices, sub-devices, modules and so forth. Can you explain this segmentation and how it can be extended?

Yongjing: oneM2M may extend the SDT model and align this with the W3C work

Andreas: the module class is a wrapper for a given set of capabiltiies

We modelling a single type of device, not relationships between devices

Martin_Bauer: we wanted to avoid recursive models for devices

We may look at defining a group of things as a new thing in its own right

Darko: did you have domain experts to guide your model specification?

Yongjing: we had experts from a variety of companies that provide devices and services.

We will only start work if we’re confident we have the experts available to guide it

<Zakim> kaz, you wanted to ask about how to validate TD using HAIM

Kaz: interested in how to use HAIM as a semantic model for validating thing descriptions

Yongjing: we could explore this for a plugfest

But who is volunteering to drive this effort?

Perhaps Fujitsu?

Kaz: will you be at TPAC next month?

Yongjing: sorry, no

oneM2M Base Ontology

Martin presents slides by Joerg Swetina who is now on vacation

OneM2M release 1 was agnostic on meaning of data carried. Release 2 added support for adding metadata, and release 3 will include specific metadata models

We are interested in syntactic interoperability and semantic interoperability

The latter involves annotation of M2M data, abstraction from specific technologies and data brokering

We support ontology references and semantic descriptors

We want to support abstracion, to enable semantic queries and semantic based mash-ups

Martin presents the oneM2M base ontology

Devices are things you can interact with

We try to distinguish between machine/tech dependent concepts and human meaningful concepts.

The base ontology can be sub-classed for mapping to other ontologies

<mkovatsc> Kaz, your question on this part?

<mkovatsc> okay :)

Darko: how do you use the semantic description? Does it describe how to interact with a device?

<McCool> to Zakim, not used to IRC, will have to look at manual or something

<McCool> to Zakim, never mind for now

<michael> can we imagine how TD could be used to bridge a oneM2M system to a non-oneM2M system?

<Zakim> kaz, you wanted to ask if the oneM2M Base Ontology is RDF-based

Kaz: what is the concrete representation are you using and is it based on RDF?

Martin: We use RDF triples serialize as XML

we may look at JSON-LD in the future

Kaz: is the combination of RDF and SPARQL powerful enough for you?

Matthias: what motivated the distinction between machine and human concepts? If the W3C WoT TD we didn’t see a need for that.

Martin: the idea is to separate the functionality from the technical implementation

Matthias: we are out of time so we won’t talk about protocol bindings now. Should we now discuss next steps?

Next steps

Martin: if we want a large group from both W3C and oneM2M we will probably be stuck with a teleconference, what do others think?

Yongjing: I was hoping to create some formal liaison arrangement

At the group to group level we can continue to exchange ideas

Matthias: the next call will have to be after TPAC, i.e. late September or October

<kaz> TPAC page (Sep. 19-23)

Yongjing: we’re having a meeting in Sophia Antipolis in October

If any of you want to attend please email me

We have regular phone calls for the semantic group on Mondays

<kaz> oneM2M at ETSI

Our (=oneM2M) next telecon is September 5 at 14:30 CEST and is 90 mins long

<kaz> W3C Liaison page

Yongjing will be the contact on behalf of oneM2M semantics group, at least for now

<McCool> Kobe, I can do that...

Matthias brings the meeting to a close

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2016/08/31 14:01:45 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.144  of Date: 2015/11/17 08:39:34  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/preent+ Tim_Carey//
Succeeded: s/Darko(?)/Darko/
Succeeded: s/q_ Michael_Koster//
Succeeded: s/Tim Carey:/Tim_Carey:/
Succeeded: s/Martin Bauer:/Martin_Bauer:/
Succeeded: s/is SPARQL well received?/is the combination of RDF and SPARQL powerful enough for you?/
Succeeded: s/The next telecon/Our (=oneM2M) next telecon/
Found ScribeNick: dsr
Inferring Scribes: dsr

WARNING: Replacing previous Present list. (Old list: Kaz_Ashimura, Dave_Raggett, Harada_Kei, Andreas, User2, User4, Danh_le_Phuoc, Harada_Kei(again), Klaus, Martin_Bauer, Nicolas_Seydoux, Peter_Kim, Soumya_Kanti_Datta, Uday_Davuluru, Michael_McCool, DanhLePhuoc, Lindsay_Frost, Michael_Koster, Daniel_Peintner, Darko_Anicic, Fenger_Zhang, Framl+Reiscj, Istvan_Lajtos, Miao_Jiang, Shelby_Kiewel, Wolfgang_Granzow, Burhanettin_Nacar, Dongjoo_Kim, Gizem, Yingying_Chen, Matthias_Kovatsch, Victor_Kueh, DarkoAnicic, Johannes_Hund, soumya)
Use 'Present+ ... ' if you meant to add people without replacing the list,
such as: <dbooth> Present+ Kaz_Ashimura, Dave_Raggett, Harada_Kei, Andreas, User2, User4, Danh_le_Phuoc, Harada_Kei(again), Klaus, Martin_Bauer, Nicolas_Seydoux, Peter_Kim, Soumya_Kanti_Datta

Present: Kaz_Ashimura Dave_Raggett Harada_Kei Andreas User2 User4 Danh_le_Phuoc Harada_Kei(again) Klaus Martin_Bauer Nicolas_Seydoux Peter_Kim Soumya_Kanti_Datta Takuki_Kamiya Tim_Carey Victor_Charpenay Jaeho_Lee Omar_Elloumi SeungMyeong_Jeong Thomas_Kessler Masato_Ohura Antonella Ian_Blair Katsuyoshi_Naka Kazuaki_Nimura Mengxuan_Zhao Wonsuk_Lee Xueqin Yongjing_Zhang Catalina_Mladin Chonggang_Wang Hiroyuki_Maeomichi Carmelo_Zaccone
Agenda: https://www.w3.org/WoT/IG/wiki/IG_WebConf#Agenda_of_next_WoT_IG_WebConf:_31_August_2016
Got date from IRC log name: 31 Aug 2016
Guessing minutes URL: http://www.w3.org/2016/08/31-wot-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]