scribenick: sebastian
start with agenda overview
start to check the minutes from the last time
scribenick: kaz
Sebastian: Samuel Singapogu from Schneider Electroric is an invited guest today
... he is aware of the W3C Patent Policy
<kaz> W3C Patent Policy
Samuel: ok
scribenick: sebastian
<kaz> May-22
any objections to make the minutes public?
no objections
<kaz> @@@slides tbd
Michael Koster shows some slide from the iotschema.org meeting
Koster: spend time on the ODM
meta-model
... how to bring Thing class to iotschema.org
Sebastian: its about ODM classes to thing class in iotschema.org
Koster: yes, iotschema.org has only
capabilities so far
... iotschema model and OneDM looks quite the same
Lagally: if you have the IoTThing can you also build composition of Things?
Koster: yes, you can do it. Explains an outlet example
Lagally: how to handle name conflicts? Do you have namespces?
Koster: ODM uses JSON Pointer
... for the outlet example can be diffrent
... you can use protocol bingings with different TDs
Sebastian: there is also namespaces in the ODM, right?
Koster: yes, namespaces are used to
seperate the orign
... to identify a particular affordance we would use JSON
Pointer
... that is the different from RDF
... there is a proposal how path can be constructed in RDF
Victor: name conflicts are
resolved by there URI
... in iotschema.org you can assign more then one class
... this would be my proposal what a user can do to use more
then one class
Koster: that would also work
fine
... we can use this in the upcoming PlugFest
Sebastian: there are some plans to introduce JSON Pointer in the TD as well
Koster: what you want to do with it?
Sebastian: point to global data model definition
Koster: TDT is the RDF definition of
OneDM
... device manufactor prefer JSON definition
... all the IoT platfomrs like Microsoft and Vorto using the
properties, actions, event style
... we need one RDF representation of the ODM and TDT would be
a good candidate
Sebastian: so, the ODM is in JSON, TDT is the candidate to have ODM in RDF and will be managed by iotschema.org?
Koster: yes, that is a way. we do not need different working groups to working on the same thing.
Lagally: if want to create type annotation whould you use iotschema.org or something else?
Koster: lets use iotschema.org
Sebastian: support open capabilty
definition such as iotschema.org to avoid multiple
definitions
... what are the next steps in iotschema.org?
Koster: we want to extend it and
involved Dan
... clearify the hosting iot.schema.org or iotschema.org
<mlagally_> McCool comments that in a scenario where different companies create types/templates the use of names is not sufficient to avoid conflicts. Unique ids are required for this purpose, one common way of solving this problem it is reverse domain names (as they are used in the Java classpath). Thing description templates and thing descriptions need to take this problem into account.
samuel singapogu from schneider electrics gives an introduction
Samuel: works in R&D
department in the area digital building
... works in data modelling
... is there more information about the presenation about
iotschema.org?
... and other question, what is the relation to SAREF?
Koster: I will send you information
about ODM
... we work with Maria Poveda to work on the onthologies
... SAREF brings a lot features that are related to ODM
... we should more discover this in the PlugFest
<victor> here is a collection of TDs annotated with SAREF: https://www.vcharpenay.link/talks/td-sem-interop.html
check issue https://github.com/w3c/wot-thing-description/issues/905
<victor> here is the paper describing what we've done with this dataset (written by Sebastian and myself): https://www.vcharpenay.link/publications/2020-eswc.pdf
next issue: https://github.com/w3c/wot-thing-description/issues/902
is a new issue
Ege: its about the location of the TDs. how to share this information to the others (the clients)
Sebastian: question about the format. is it TD or the link format directly?
Ege: its the td
Koster: we can use different rel types like host
Victor: there is a URI to identify the TD and there is a ID to identify the physical object. that is a different
next issue: https://github.com/w3c/wot-thing-description/issues/903
<Ege> https://github.com/w3c/wot-thing-description/pull/907
next: it's about the dynamic control of hypermedia
https://github.com/w3c/wot-thing-description/tree/master/proposals/hypermedia-control
question: is it possible that a client can cancel an action which it not initated?
Koster: yes , this is possbile
<Ege> https://github.com/w3c/wot-thing-description/issues/899#issuecomment-635642205
Sebastian: Ben mentioned that we should
not make TD that complicated as you can find for OpenAPI. He
suggest to use an own subprotocol that implements the
hypermedia control
... proposal to evaulate one of the hypermedia approaches from
Victor and Ege in the PlugFest. Decide after if it make sense
to include such a feature in the TD or not.
Klaus: are both approaches static?
Sebastian: ege's version is more static, Victor's dynamic
Klaus: to get an overview of
active actions you need dynamic TDs
... we do a mini version of OpenAPI here
... would it make sense to take TDT and use OpenAPI for the
interaction?
<Zakim> kaz, you wanted to point out we're at the top of the hour
<kaz> [adjourned]