<kaz> scribenick: ege
<kaz> TPAC registration
<kaz> Accommodation
McCool: can taki attend the workshop for data models
<kaz> workshop cfp page (12-13 Sep., Palo Alto, US)
Taki: I haven't booked flights so
I can attend this workshop and fly from there
... physically I am ok
McCool: a proposal together?
Kaz: like for our workshop, a statement
Taki: then I can submit today
<kaz> application form
McCool: summarizing what we have done
McCool: there is a wishi meeting and
hackathon
... michael koster is participating
... any other conference that is relevant?
Koster: friday 19th we are having a meeting with OMA
<mjkoster> this is the agenda for the IETF T2TRG/OMA meeting on Friday the 19th
<mjkoster> https://github.com/t2trg/2019-07-oma
McCool: I kind of invited myself to that
Daniel: googling td on google shows up an old TD proposal that is still marked with a recent date
Daniel: it is called JSON Web Thing Description
McCool: this shouldn't even be there, it looks like an official document
Ege: it is not actually published at this date
Kaz: it's just a date respec generates, isn't it?
McCool: looks like it
Daniel: it is in our repo so we can simply remove it
Kaz: we have to tell the Authors/Editors of those old documents gently before removing the documents to remove it
McCool: maybe we can put in an archive subdirectory
<kaz> proposal directory
McCool: there are a lot of documents in here
Kaz: maybe just change the "proposal" folder name to "archive"
<kaz> old architecture directory
(mm creates a directory called ARCHIVE under proposals)
Lagally: there is a folder called architecture that should not be in the wot main repo
Koster: I think it is better to keep the folder names but simply moving everything under archive
McCool: (works on cleaning up the main directory)
Kaz: if it is ok for Lagally, we can
move the old "architecture" directory under "proposals"
first
... and then ask all the editors of old documents about if it's
ok to change the name of "proposals" to "archive"
<Zakim> dape, you wanted to assign someone to check github for outdated information (to put into ARCHIVE folder)
McCool: writing an email to the
list
... need to look at what is old, like 2 years
McCool: 6-month exension to WG has been approved (till Dec 31)
(Kaz is working on the AC announcement.)
<kaz> Wendy's comment
McCool: giving a wiki site for
exploration
... maybe don't need for clarifications on all the bullet points?
... some things are intentionally vague?
<kaz> Proposed IG Charter
McCool: (looks at the PRs by
wendy)
... let's merge the charted PR
<kaz> (some minor conflict and McCool will look into it later)
Kaz: we'd like clarification from alan bird
<kaz> Alan's comment for IG Charter (Member-only)
McCool: we have proposed a clear
sentence for this comment to be added
... I will ping him and ask him
Kaz: any joint meeting ideas?
McCool: JSON-LD and accesibility
would be good
... for Monday, Sep 16 (18:30-22:30)
... I can go to dev meetup
... other attendees can write their names here but you have to
register first
Lagally: reviews from companies and
individuals show some issues
... you have now a formal way of describing devices
... each time a new class of device comes, we can get
un-interoperability
... so we need profiles that can guarantee usability of the
device
... so we need some constraints
... having only one form
... having string or array of strings, causes an overhead
... constraint on objects, not nested
... not permitting security on lower levels, only one at the
top level
... constraints on protocol, having only http
McCool: what reduces complexity and
what reduces interoperability
... so we should have two categories
Lagally: so that constrained devices
will be easier to support
... we can have multiple profiles
McCool: we can look at the test results and focus on features that are implemented by 3+ implementations
Lagally: but what is there as an example to validate the spec and what is an actual product?
McCool: we can also add to check what uses HTTP or other profiles
Ege: mozilla also talked about having action semantics (a queue for running actions etc.) so maybe we should look into specific behavior of interactions
Lagally: interesting point to look
into
... where would a new implementer find a guide to do a first
implementation
... high level constaints
... (talks about the bullet points of the last slide)
... I was thinking of specifying a default profile
... the proposal is to have a very easy to implement
profile
... I am working on a list of constraints
McCool: I would keep it simple
Lagally: the difficult part would be to define the simple one, a complex one is easier to define but something that can satisfy 80% use cases
<Zakim> dape, you wanted to think about ways to identify capabablities.. profile handshake?
Daniel: handshake definition
McCool: Lagally is proposing the simplest
possible
... there will be very probably multiple profiles, smart home
and industrial for example
Lagally: you can do anything, just that if you stick to the core profile you have a guarantee of wider usability
McCool: we need to clarify what are the allowed extensions
Lagally: I expect to have a draft within a week
[Thing Description]
<kaz> Jul-5 TD minutes (member-only)
Ege: protocol bindings will be with TD calls and I will be co-editor of the spec
[Archtiecture]
<kaz> Jul-4 Arch minutes (member-only)
Lagally: architecture - we read the
comments of David Baron
... device management is out of scope
... we had some useful comments
... like missing links for the external links
... they also didn't understand what will happen if you parse
JSON or JSON-LD
<inserted> kaz: note that I sent another reminder to TAG and mentioned it might be going to make sense to have a joint call with them to clarify their questions.
McCool: IETF is happening in two weeks
Lagally: I would organize the call
next week then
... saying that this would be a
deadline for us since the week after there will be the IETF
meeting
Kaz: Daniel Applequist, Yves Lafon and David
Baron should be involved
... we should clarify their viewpoints as TAG as a whole
[Scripting API]
<kaz> (no call on July 8 but updates from Zoltan)
scribenick: egekorkan
Zoltan: so there
are some projects that are already using web of things
... and they need some help
<zkis> https://www.sofie-iot.eu/
McCool: we have extensions
... but how do you do in the scripting api
Zoltan: scripting is orthogonal
[Security]
McCool: security is handled out of hand
Zoltan: in payments it is transparent, it is happening in the backend
McCool: 5 mins break before the plugfest
Lagally: let's stay on this call
<kaz> [adjourned]