Web of Things IG - TF-AP/TF-TD joint meeting

12 Aug 2015

See also: IRC log


kaz, sebastian, arne, claes, daniel, darko, johannes, kajimoto, nimura, markus, taki, matsukura, Claes_Nilsson
Johannes, Sebastian


<jhund> please come here for the joint call of TD and AP

<scribe> scribenick: kaz

sebastian: first call after the f2f meeting
... thing description and application protocol
... we'll clarify which call will happen next week, etc.
... report from application protocol TF

TF-AP report

johannes: first question is report from the Sunnyvale meeting
... and report by Darko
... would give report from the Sunnyvale meeting
... presentation about the outcome
... had a breakout session
... also joint discussion with TD/SP
... about web of things model

(shows presentation https://www.w3.org/WoT/IG/wiki/images/8/8c/TF-AP_breakout_report_consensus.pdf)

johannes: architecture model
... technology landscape
... lifecycle states
... abstract resource model
... identified iteration
... still need further refinement

(picture of three WoT Servient: p4)

johannes: use Web Protocol Client to access remote WoT Servient
... direct access or via some Web service
... existing domain specific protocols may be used like bluetooth
... need some refinement

(Points to address)

johannes: difference between protocols
... orange vs blue
... constrained device
... what is the minimal servient
... one servient hosting several virtual instances
... servecies in the cloud

(Technology landscape of TF-AP)

johannes: protocols, resource models and api patterns

(Lifecycle states of a WoT Servient)

johannes: discussion initiated by Kajimoto-san
... offline/standalone
... registered/paired
... activated/connected
... important for security viewpoint as well

(Abstract Resource Model)

(Protocol-agnostic thing model for web things)

johannes: protocol agnostic way to express things
... would handle HTTP, CoAP, etc., at once
... meta data should be also handled

(Runtime Properties)

johannes: create interaction
... dynamic operations
... read,write, subscribe, observe
... read data at some specific point


johannes: invocable action on the physical thing
... issue a state change?
... enables us atomic change of multiple resources

(Next steps)

johannes: need to complete and transfer the technical landscape

sebastian: some gaps? could you please repeat the last part?

johannes: tech landscape from the wiki
... and evaluation models
... to discover what would be the best way
... define a model based on the use cases
... that is my report from the Sunnyvale meeting
... anybody from AP-TF is welcome to comment

sebastian: link for the use cases?

-> http://w3c.github.io/wot/tree/master/plugfest

-> https://github.com/w3c/wot/tree/master/plugfest Plugfest page

johannes: there is a github as above

sebastian: making examples based on use cases described here

johannes: would have more consistent plugfest

kaz: how could we put together all the use cases that are already on the WoT wiki pages

johannes: agree with we should gather all the use cases
... though not sure how to use them for the plugfest scenario
... would be pretty obvious on home electronics use cases here

sebastian: some of the use cases fitting quite well with this plugfest scenario
... very nice scenario
... home automation

kaz: thinks it would be nice to concentrate on home automation use cases for plugfest
... and probably some of the TF-AP participants should be able to make contribution for that

johannes: before TPAC demonstration, we need to consider what to be modeled

kajimoto: any idea to include any thing description into this?
... maybe good sample for that purpose

johannes: good to provide some sample
... TF-TD could provide some notation for that

sebastian: very good experience
... to decide what the basic model would be
... and see what is missing

kaz: yeah, so we could think about some initial small profile of the Thing Description for this Plugfest use case

sebastian: exactly

johannes: we just started the discussion
... any more comments/questions on the Plugfest?
... we can have discussion on Github and the ML

sebastian: yes
... btw, who would take the slot next week?

johannes: next week for TD or AP?
... myself don't care :)

sebastian: would like to have TD next week unless there is any specific opinion

johannes: ok
... next Wed. 19th should be TD call
... and AP call on 26th
... if there are no objections

sebastian: ok
... next week we'll have a TD call on 19th
... and AP on 26th
... also on Thursdays we'll have SP/DI calls
... we'll do that until TPAC

johannes: need to leave now

sebastian: would take over

TF-TD report

sebastian: shows the agenda of the Sunnyvale meeting
... outcome from the breakout

<Sebastian> https://www.w3.org/WoT/IG/wiki/images/d/d1/Breakout_report_sunnyvale.pdf

sebastian: TD model
... minimal set of vocabulary
... data type
... JSON-LD examples

(TD Model)

sebastian: metadata component
... name of the things, product ID, etc.
... also includes what kind of protocols are supported
... encoding information
... EXI, JSON, etc.
... data and data type
... reuse existing data type
... refer to data types defined here
... interaction model pattern
... property, action and event
... that is the basic structure of the proposed Thing Description

darko: information on APIs?
... accessible by get

sebastian: this is a basic model
... what kind of property has which charasteristics depends on vocabulary
... what we're doing here is very abstract modelling
... if we have a concrete protocols, we need to think about how to map the protocol to this model
... would talk about a bit more concrete categories
... describe the components more precisely

(Minimal Vocabulary Set)

sebastian: metadata: name, protocols and encoding
... which has to be written
... could use some specific vendor's protocol
... could reuse existing ontologies like SSN
... data: xsd convention
... definition by XSD
... able to define data types and data access manners
... XSD is very powerful for that purpose
... and well-known property: name, input/output, writable
... need to clarify by TPAC

(Data Type Definition)

sebastian: rely on XSD data types
... define suitable subset
... proposals to be discussed during the next meetings

(Examples in JSON-LD)

sebastian: not finished yet

(JSON-LD (Snippet))

sebastian: IETF thing2thing meeting
... will continue the discussion on present sensor, switch, etc.

-> https://github.com/w3c/wot/tree/master/plugfest Plugfest scenario

sebastiona: see "Things" section of the above plugfest scenario

(For the next Web Meetings)

sebastian: will check relevant security aspects
... volunteers for participation
... demonstrators
... that's it
... happy with the outcome from the Sunnyvale meeting
... questions?

kaz: what do you think about the possible collaboration between the TF-TD and the TF-AP?
... could both the TFs work together for the Plugfest demo at TPAC 2015?

sebastian: unfortunately, couldn't attend the IETF meeting in Plague itself
... not full potential of TD for the demo but could provide some descriptions for that purpose
... would look more protocol implementations

kaz: maybe we might want to have a joint call once a month, etc.

sebastian: might be a good idea
... we'll share a same slot in any case

Report on Spatial Data by Darko

darko: would inform on the Spatial Data group
... possible collaboration
... our focus is defining minimum vocabulary for Thing Description
... we might get help from other groups who work on related topics
... Spatial Data group works with the OGC
... related to vocabulary/ontology work
... trying to make ontology standard
... standard vocabulary for sensors
... our minimal vocabulary can be mapped with SSN ontology
... data provided by other sensors
... actuators as well
... semantic description for actuators is missing withing W3C
... Spatial Data group is working on actuators as well
... would input from us WoT IG
... requirements on vocabulary
... if we could provide requirements on semantic model for actuators
... would be great
... Carry Tailor (?) would input from us
... would have a joint session with us
... during TPAC

sebastian: interesting
... ontology outside for sensors, etc.
... what should we do if we want to use that kind of ontology with Thing Description?
... how to handle company specific ontology?

darko: a few examples describing sensors
... discovery on the basis of SSN ontology
... described by the Thing Description language?
... what the traditional aspect is like
... what is the capability of sensors?
... depending on the conditions
... important aspects for applications
... would be good to work with other groups who are taking cares
... query devices based on the capability descriptions

sebastian: tx
... think we should invite them to our calls
... to see how does it fit with our model

arne: make sense to find thing description model
... new version of SSN would include actuator capability

sebastian: building concepts based on the new version of SSN?

arne: defining a light-weight profile based on that

sebastian: would talk with related groups
... and let them know what kind of problems we have
... and see if we could simply reuse their mechanisms, e.g., SSN

darko: SSN is widely used
... would be better to see if SSN description suffices
... not actuators yet

sebastian: we're getting out of time
... next week we'll start our TD call at this time
... and would have joint meetings like today

[ adjourned ]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/08/12 14:35:57 $