<kaz> scribenick: taki
<inserted> Danh's slides
Darko: We continue to discuss
plugfest preparation.
... work from Danh and victor.
Danh: Have been discussing query
in thing directory.
... sparql has some problems.
... two solutions. using JSON-LD frame.
... By the end of this week, we plan to make PR.
... query for thing that provides temperatures by
filtering
... frame is simple. create JSON-LD frame query and send
it.
... to end point.
... and it returns the partial document that contains data of
interest.
... query is in JSON format.
... you can modify the content on the fly based on
templates.
... you can create on-the-fly reversed property.
... also implementing graphQL
... query becomes even shorter with graphQL
... GraphQL uses Shacl for validation.
Kaz: Can you send the presentation slide, Danh?
Danh: ok.
Koster: filtered lookup, I have
been advocating. This can be a solution. great.
... GraphQL works under the hood.
Danh: We can also use this to
other applications.
... JSON-LD 1.1 frame, I believe is in the process of
standardization.
Koster: declarative filter is useful.
Darko: you don't need
conversion.
... This is better.
... You can put recipe directly into query.
<DanhLePhuoc> https://www.topquadrant.com/graphql/graphql-shacl.html
<victor> https://www.topquadrant.com/graphql/directives.html
<victor> that's it, right?
<DanhLePhuoc> @Victor: yes, it is
Danh: I will present more on Open Day, including demo.
Darko: shapes validate data.
It automatically converted to sparql query.
Darko: what is the
timeline?
... will it be ready by plugfest?
Danh: JSON-frame work is on the
way, this weekend, hopefully can make PR.
... Next week, people should be able to try it out.
... I will also do test.
<Zakim> kaz, you wanted to ask if this information should be included in the PlugFest preparation.md file (question after the technical discussion)
Kaz: If this is expected to be
provided by Plugfest, we should provide some description. But
it seems this is some of more advanced features.
... some WoT members are interested in this feature.
Darko: In the past, we created some slides (step1, step2, etc) to explain sparql-based query.
Danh: I will follow the style.
Kaz: I should ask Koster (plugfest moderator) if it is appropriate.
Koster: I agree with adding
it.
... How to use it, then how it is implemented.
... Whatever we add, let's provide description.
Danh: You can use the same thing directory.
<inserted> Aparna's slides
Darko: let's proceed to next
agenda item. Feature of Interest.
... Aparna, can you show the current proposal?
Aparna: iotschema.org, FoI
pattern is discussed.
... FoI pattern came from SSN SOSA
... FeatureOfInterest class.
... capability is associated with featureofinterest
... property is a property of FoI, etc.
... Dataschema has actionOn association to Action.
Aparna showing TD example.
Aparna: pumpstatus "isPropertyOf"
... a FoI
... DataSchema, description of data. Now it is a bit unclear
how to associate with FoI.
<Zakim> kaz, you wanted to ask Aparna (as well :) to provide the slides
Koster: I created an example too.
It is exactly the same.
... I had the same question too.
... we are in the same stage.
... why we added triples instead of links. it is more
appropriate.
... we need to add something to context to enable dataschema
FoI.
... extension point, we will need to consider.
... For now, directly adding to TD.
... we can also add it into iotschema vocabulary.
Darko: with FoI, we are making a need for extra ontology to derescribe your system. this is interesting.
Aparna: festoPA:pipe2 is it concept or instance
Koster: It is instance
Darko: there are two tanks and
two pipes in the TD we are making for plugfest.
... festo-1, etc.
Danh: new pattern FoI, will it be iotschema or TD vocabulary?
Koster: iotschema vocabulary.
Darko: you can see the classes are in iot namespace.
Danh: FoI is new ? or are we just referencing SSN SOSA?
Darko: FoI is aligned with
SSN.
... we did not take it as a whole from SSN.
... we took a part of it.
... event and observable relationship, for example. We adapted
to our pattern.
Koster: We integrated it into our capability pattern.
Danh: importing SSN as a whole
may be too big. So it sounds like a good solution.
... we will need to pay more attention to alignment.
Danh and Koster discussing tooling including wol, RDFS, SHACL, etc.
Kaz: We need more discussion during F2F.
Darko: We will provide TD for
FestoPA.
... Can query TD based on FoI.
... using areas, doors, pipes.
... I plan to attend the F2F remotely.
Koster: I will also add FoI in
TD.
... we should not block people from using it.
Kaz: It is better to document in preparation.md as experimental. and if it's difficult, would be good to document it within the result.md later.
Koster: we should inform the IG on how to use it.
Darko: I will describe in preparation.md.
Koster: Automotive use case is also interesting. I will contact Benjamin and Soumya.
Darko: This is the last meeting
before F2F.
... Danh and I will add some description in preparation.md.
Kaz: Danh and Aparna, please send in slides.
<kaz> [adjourned]