<scribe> scribenick: kaz

iotschema

Koster's slides

Darko: summary from Koster?

Koster: had some slide set
... iotschema repo
... let me go with an updated one
... [HTML Website Protytype]
... [WISHI Semantic Interop]
... used a lot of WoT infrastructure
... first time put everything together
... LWM2M, OCF, vendor speciric ones
... Thing Directory, TD with semantic annotation from the iot.schema.org prototype definition
... hypermedia control
... some activities on semantic interoperability
... WISHI was a workshop held last year
... [W3C Web of THings Plugfest]
... and our own plugfest in Prague
... semantic annotation, protoco bindig and interaction through proxy servients
... included automotive domain by Eurecom and BMW
... covering GENIVI VSS automotive definition
... [Learning from Plugfests]
... directory-based discovery using semantic annoation and queries
... how to explain that to people
... people need to learn the basic concept
... not a lot to ask for
... need user guide, whitepaper, etc.
... people expect a lot of definitions
... looking at automotive and smart buildings
... realization of feature of interest
... need to add a lot of extensions
... concept is already available in haystack
... automotive describes basically automobiles
... a set of definitions
... property types and relations
... [What is new in iot.schema.org]

Darko: don't need to go into the detail at the moment
... what is coming to iot.schema.org

Koster: we added some automatic events, actions, etc.
... [Working Plan]
... conclusions from W3C WoT Meeting
... integrate feature of interest
... [iotschema Capability Pattern]
... clarified some pattern
... [RDF Shape Constraints]
... use RDF Shape Constraints to define data shapes
... [Next Steps]
... get the WoT CG Charter approved
... update the base model to include dataItem and Feature of Interest concepts
... define dataItem constraint mechanism
...

Darko: what we had so far
... PlugFest in Prague
... annotated available devices
... how discovery could work
... also played with automated generation of TD
... also showed how application can be generated by recipe
... discovered ingredients of recipe
... have been the points so far

documentation and plugfest preparation

Kaz: are you planning to document the achievement?

Darko: yeah
... as part of the plugfest report already

<DarkoAnicic> Siemens contribution at the last PlugFest in Prague: https://github.com/w3c/wot/blob/master/plugfest/2018-prague/result-siemens.md

Darko: 3 parts there
... recipe and ingredients
... but this was not related to T2T
... we should write this report from our TF
... as a CG Note
... should be there

Kaz: some more details?

Darko: yes
... and new content as well
... any idea about semantic work?
... contribution to the plugfest

McCool: one thing is security metadata
... could you standardize vocabulary for that purpose?

Darko: ok
... how shall we contribute?
... e.g., within which TF?

McCool: not sure
... additional role
... scripting configuration, etc.
... maybe something like semantic annotation
... TD vocabulary by TD TF
... worthwhile to look at how to model that
... security concepts are up there

Koster: work already done for access control, etc.

McCool: ontology of NEST partly covers that
... kind of open area still

Koster: vocabulary along with the HTTP

McCool: requires some study

Danh: want to look for vocabulary for access control?
... access policy as well?

McCool: maybe
... certain configuration
... need to think about concrete use cases
... need some standard vocabulary

Danh: would be interesting
... based on the basic vocabulary
... accessing policy together with the description
... good way to do that is documenting to specify vocabulary

McCool: maybe you can provide your ideas by email
... and we can add that to the security repo

Darko: tx for your input
... how to proceed then?
... security tf?

McCool: OCF resource types to see capability
... need exercise
... for reversible OCF meta bridge
... what kind of information do you need?

Koster: good activity
... and great if someone could create mapping pattern
... whatever we need to add
... you can do that top down from RDF

McCool: my OCF devices are now visible using the bridge
... my metadata bridge would be live during the weekend

Koster: TD annotations representing OCF devices
... OCF devices are good starting point
... and would say Zigbee devices are also
... a lot more actions, properties for each definition

McCool: interesting trials there
... oneM2M to OCF, etc.
... having 2 different ecosystems would be good exercise

Darko: wondering about oneiota

<DarkoAnicic> oneiota: https://oneiota.org/documents?filter%5Bmedia_type%5D=application%2Framl%2Byaml

McCool: we can digest it automatically
... resource type in OCF corresponding to general resource type
... manually annotated first
... take union of capabilities

McCool: trying what I can

Darko: some of the existing capability effect

McCool: the question is abstractly that's correct
... but need to clarify concrete vocabulary

Koster: pretty much one-to-one mapping
... how would that work with some concrete vocabulary

Darko: true
... we should consider some more resources
... in addition to OCF, etc.
... iotschema should be a normalized layer
... for existing semantics
... could be translated to existing semantics

McCool: script is normalizing mechanism
... we need to do more than one ecosystem

Koster: Ericsson guys are working on IPSO
... bringing a few people who could work for other ecosystems
... go to see what would be abstraction
... sounds like there are 3 at least
... smartthings as well
... zigbee, ocf, ipso as well

Darko: ok
... you're advocating that
... would be one point
... on the other hand, for WoT plugfest
... annotation for existing descriptions
... iotschema to cover what people need
... what could we demonstrate?

Koster: we need to work on
... we could increase automation
... also work on feature of interest
... giving people more interest
... need more examples
... to make the learning curve less steep

Darko: right
... more online things
... to allow a number of participants to provide ideas

Koster: good point
... using the Internet remotely

Danh: showcase how to use TD for semantic sensor net, etc.
... using TD directory
... not only for our plugfest

Kaz: maybe related our discussion on online testing framework?

Darko: partly

McCool: continuous validation, etc.

<McCool> https://github.com/w3c/wot/blob/master/testing/online.md

McCool: not quite done yet
... secure HTTPS framework
... will work on authentication
... update to track latest update of TD
... generate a script and test device
... possibly a couple of hundred devices

Koster: simulated devices are fine
... semantic annotation on the directory
... high-level of security requirement might be limited

McCool: simply basic authentication first

Koster: yeah

McCool: doing similar device to OCF
... want to figure out automated translation
... want to do that exhaustedly

Koster: mostly thing directory

McCool: will give me a SPARQL endpoint as well
... e.g., someone invoke a thing directory and I can use that
... once the discovery script api is standardized, new prototype should follow that

Darko: really interesting
... and good starting point
... we need to make it easier for people to use
... e.g., add this kind of features to the Thing Directory
... for the next TF call, focus on providing some ideas on semantics for online scenario

McCool: we need to have a killer application (or two)

Koster: we also need some more use cases
... some clear use cases in my mind
... some scenario which work with concrete use cases
... really like the idea of recipe but still need use cases

Darko: ok
... we'll have the next call in 2 weeks
... this kind of distributed environment would be great
... what could we do with these multiple sensors, etc.
... some other point might be need to scale iotschema for larger vocabularies
... Korea is far away from Europe
... talk to you in 2 weeks

[adjourned]