<benfrancis> Note to add to the meeting notes, would like to announce a test instance of the Things Gateway by Mozilla, for interoperability testing https://lists.w3.org/Archives/Public/public-wot-ig/2018May/0040.html
<mkovatsc> https://www.w3.org/WoT/IG/wiki/Main_WoT_WebConf#30_May_2018
<kaz> scribenick: mlagally
testing TF - new timeslot - Monday 7am PDT / 10am EDT, 3PM UK, 4 PM EST, ..., 11PM Japan
Dave: reference is UTC?
McCool: yes, UTC keeps calendars
Kaz: AC rep discussion is still ongoing - please note that most of the W3C telcos are allocated based on US Eastern time (and those would be effected). so if we want to use UTC, everybody has to change their calendars
Ben: Mozilla test gateway
is available for interop testing
... not yet implementing official W3C spec - converging
McCool: please add link to "testing" page
Matthias: timeline for convergence?
ben: dpeends on w3c spec
evolution. It is very fluent at the moment
... expect to make new round of comments on latest TD spec
McCool: need to validate implementations against the standard -
Matthias: if it is not implementing the current spec, there should be a disclaimer
ben: feel free to link to our spec on our website - this is pretty close to the W3C spec
McCool: there should be links to validation reports
<kaz> plugfest reports
McCool: intel documentation is almost done - will probably do a small cleanup - keep in edit for the moment
Koster: still need to update the report
Darko: I did my part, but also need to add documentation about TD generator based on IoT schema as requested by Kaz
<mkovatsc> Korea f2f wiki
Matthias: Please register at the
Wiki page and fill out the questionnaire
... open day: several emails to local institutions and
universities were sent. Kaz will reach out to Samsung and LG AC
rep
<mkovatsc> https://www.w3.org/WoT/IG/wiki/F2F_meeting,_30_June-5_July_2018,_Bundang,_Korea#Input
Matthias: please update input
section on Wiki page if you have topics
... plugfest calls will be after this slot to discuss
details
Kaz: hotels and flights are filling up, please book asap
McCool: when do we end on last day?
Matthias: 4pm
Sebastian: <presents previous W3C
working draft and new W3C Editors draft>
... final vocabularies will be consolidated this Friday
... Taki will moderate the call on Friday
McCool: security definitions are wrong - need to update
Sebastian: security is the last unstable thing, everything else is quite stable
McCool: we could also use one additional week to stabilize security
Koster: security and form are extension points. If you dig further there will be additional vocabularies, especially if new protocols are included
<inserted> https://www.w3.org/TR/2018/WD-wot-thing-description-20180405/
<sebastian> https://w3c.github.io/wot-thing-description/#overview
Sebastian: "interaction" term was
removed
... new terms are "properties", "actions" and "event"
... we still have the interactionPattern for simplification
... small change on dataSchema - renamed to "input" and
"output"
... we are 100% compatible with JSON schema (supporting a
subset)
... we are also compatible with JSON-LD 1.1
... previous version had heavy array usage - we still can keep
our goal of JSON style + semantic annotations and existing
standards relying on RDF
... clear definition what's provided in the dataschema
... you can find a subset of JSON-schema terms
... t.b.d. if additional JSON-schema terms are needed
... you can define specific datatypes as needed
... all information is reflected in vocabulary part
... section 6 has been updated with a description of how
instance is created
... hopefully all steps are clear
... there's also a section about JSON processing
... there is no change in the basic idea, examples look the
same
... we hope everything is no fine for the next plug fest
McCool: definitions: you wanted to have name - value pairs
Matthias: this conflicts with JSON-schema. Would be an optimisation
McCool: I will just leave that
feature out of the security proposal. Less convenient, but not
a big deal
... will adopt class-structure
and create security descriptions accordingly
Koster: how to process default
context?
... if you don't include the context, there are some
assumptions
Matthias: this is how it should work. default context applies
Koster: you could also have additional context URIs
Matthias: combining contexts from
media types, default, iotschema is possible
... representation format is defined by media type. you can
select the default context from that
McCool: you can always use the default context
Sebastian: if you're not doing semantic processing you don't need it
McCool: what if I only give iotschema.org context - processing still uses default TD?
Matthias: depends on how you
process it.
... if you use JOSN-LD pipeline - use what'S there and merge
with the default context. we should improve the text to make it
more clear
Koster: on a plain JSON, can we still include @context tags?
Matthias: yes, using a plain JSON library you would do some hard-coded processing
Koster: also include @id?
Matthias: a thing always should have an identifier - id field is a masked @id identifier
Koster: will make a second written comment about subclass etc.
Matthias: We discussed ontology structure with Victor - everything is resolved quite nicely
Koster: will suggest small improvement
Koster: need to think about it
Maria: need to have another look - this seems to be coupling two models. no link from thing to interaction patterns
Matthias: layout problem?
Sebastian: we can improve
mlagally: what about webvowl diagram?
Matthias: there were some tooling issues - some additional confusing lines
Sebastian: orefer a clear model without
having to have RDF knowledge
... owls visualisation is not suitable (does not depict the
cardinality)
<mariapoveda> http://fairreviews.linkeddata.es/def/core/index.html
Maria: I put another UML
example into the chat
... we could also use this diagram
Matthias: did you do manual CSS for that?
Maria: no
Sebastian: can we add cardinality?
Maria: yes
McCool: better wait until the diagram is stable
Matthias: I will create an issue to make sure that a single diagram depicts all relevant entities
Koster: there may some vocabulary changes input / output, ...
Sebastian: we would have 3 media
types
... need to integrate into ontology
Koster: <suggests model alignments>
Sebastian: discuss on Friday
Koster: will create an issue
Maria: no link between thing
and interaction pattern
... you need to join 3 queries
Matthias: some info could be
inferred
... we could perhaps describe querying in a separate
document
Darko: interaction pattern has
not yet been described
... property, event, action need to be integrated
Matthias: there are two aspects
to the ontology: serialisation and querying
... we should separate it
<kaz> [adjourned]