See also: IRC log
<kaz> scribenick: McCool
next week no TD call
as noted in agenda
next TD call will be Friday 21 April
<scribe> new repo only for TD
suggest all issues related to TD should go in new repo
w3c/wot-thing-description
<kaz> from: https://github.com/w3c/wot/issues to: https://github.com/w3c/wot-thing-description
existing TD issues moved, eg binary description
<kaz> https://github.com/w3c/wot/issues/307
issue with not very clear description of data types
type definitions seem to be overly complicated... there is alternative proposed
however, existing format allows for semantic tagging
Sebastian points to discussion on github for details... JSON Schema interactions
have contact with JSON schema group
should like to add new item to JSON Schema definition to add semantic annotations
<kaz> issue-290
maybe makes sense to initiate proposal with JSON Schema group
would like to have concrete proposal for Osaka F2F, so we can make a decision
first step: ask for volunteers to work on a proposal
in past had a subgroup to work on data type definitions...
Daniel: proposing different approach
feel they are the experts, so...
perhaps start the discussion first in an issue
can still do a proposal, but first we should see their stand
then we can determine best way to move forward
sebastian: most important thing is to start discussion (with IETF group?)
perhaps we make a small example and start the discussion this way
andrew already has a good understanding of what we would like to have
we are not experts, but we should work together with them on this
still makes sense to pick a subgroup
Taki and Daniel ok?
Taki: so where should we start?
andrew aware of us at Santa Clara mtg, but nothing has happened so far
Sebastian: github issue would be a more formal way to move this forward (rather than the mailing list)
kaz: which group?
<dape> https://github.com/json-schema-org/json-schema-spec
can andrew simply join *this* group
Sebastian: thought we would create issue in JSON schema group
we are asking them to modify JSON, so it makes sense to make the proposal/issue on their group
kaz: we should decide what core issue is
then we should enter it on their system, then track it with an issue here that refers to the issue on the JSON Schema group
kaz; we should come up with a specific use case
that needs these issues
Sebastian: can refer also to current practices document, issues on TD
example is that we would like to integrate semantics into the data type definition itself
in addition, would be useful to have Victor in the loop
he may have some suggestions on how to do this integration
Sebastian: this is just a starting point to the discussion
Daniel: so this path means we will not change this before the Osaka meeting...
Sebastian: realistically, no time to make change to data type definitions
only have about 4-5 weeks left before F2F
need stable version 2-3 weeks before
so we need to nail down version for Osaka by end of April: 24 or 25
only about 2 weeks left for any new implementation
and... no meeting next week
probably no change in data types for Osaka meeting... but afterwards, if we can coordinate with JSON Schema group, can come up with concrete proposal for next F2F in Germany
Sebastian: for Osaka, want concrete and stable TD model
comments? Group agrees
sebastian: want clear model written up by time of F2F
some points of discussion in agenda, Maria's model
(brings up Maria's model in slide)
however... want to modify a few things
first thing, Sebastian wants to remove relative Endpoint
href can already be a relative path
other things already listed in the agenda
next point: default value definition, remove
it has impact on the data type definition, should be defined there (eg in JSON Schema)
no need to do it at the property level, can be done at the data type level
also, suggest removing the differentiation between physical and virtual thing
but causes problems with Things that have aspects of both
next, ask to change how communication metadata is defined... protocol bindings, how does this apply here?
Sebastian has asked Matthias for input here
next, Sebastian would like to change data format class to just data type; make note that we are relying on JSON Schema definition
also need to define namespaces; Maria used her own, we need a more official namespace from the W3C
we need our own namespace that is unique to TD... need to decide how to do
and we should look for things we can reuse, eg href
from other namespaces
or rdf:label for names
need a clear list what is really "new" for WoT
Sebastian: Darko, would you be willing to look in to this?
Darko: yes, willing to work with
Maria on these issues
... few more comments
... what about communication patterns and security data
here
... other thing is Data Format... Sebastian says there is
already a CR
Sebastian wants to point to schema, so would just change this, and rely on JSON Schema
Sebastian: reiterates that it
would be good to identify namespaces that are unique to
WoT
... for other stuff, should indicate what other namespaces
should be used; make a table or something
Dave already looked into this
Darko: yes, we need to talk to Dave about this
McCool: security aspects can be handled similarily to protocols
let's just add a placeholder for now
McCool: second point, adding semantics to JSON Schema useful for other applications, eg OpenAPI
Sebastian: please queue up any additional issues on the new repo
Also, want to generate new diagram to discuss in the next meeting...
then will take this model to the Osaka meeting
Sebastian: this is the plan; ok
with people? Silence... taken as assent.
... see you all in two weeks!
<kaz> [ adjourned ]