<scribe> Scribe: TK
<kaz> scribenick: taki
Kaz: Graeme Coleman is a new member.
Graeme: WoT and accessibility, I am
interested in.
... Thank you for helping me aboard.
Kajimoto: Welcome!
Kaz: Accessibility is important. Tag is interested in WoT as well.
Graeme: I will have something to present in a few weeks.
Michael Legally is introducing himself... Working for Oracle on IoT, learning interoperality with WoT. Working on other standards group. Thank you.
Kajimoto: Thank you, Michael.
Kaz: I put the summary from the presentation as PDF.
<kaz> f2f minutes
Kaz: also as HTML.
... Either version ok.
... We can finalize the minutes in the next week.
... There are many pictures in it.
... PLEASE SEND ME YOUR SLIDES TO ME.
Kajimoto: Terminology. Mozilla said, our terminology "servient" is special, which is already a common sense in our group. We should use more common term in general.
<kaz> https://www.w3.org/2017/11/04-08-wot-minutes.html#nov6-item02
Kajimoto: TD management and
publication, we need to address them, This is another
action.
... This will be discussed in TD call.
<kaz> https://www.w3.org/2017/11/04-08-wot-minutes.html#nov6-item03
Kajimoto: Nov 4 th and 5th we did Plugfest preparation at Fujitsu office.
<sebastian> sorry for my delay
Kajimoto: We achieved to
communicate between many servients as PoC across
organizations.
... We had two days WoT IG/WG meeting.
... We discussed FPWD.
... and how to use the latest PlugFest result, and explained
for newcomers.
... Please look at this minutes.
... I am one of the editors of architecture document.
... Next face to face will likely be in Prague.
<kaz> https://www.w3.org/2017/11/04-08-wot-minutes.html#nov7-item09
Kajimoto: We should upgrade FPWD by next F2F.
Kajimoto showing current plan in the document...
<kaz> Administration slides
<kaz> [[
<kaz> [Summary of the discussion on the f2f/publication planning]
<kaz> f2f planning:
<kaz> 2018-03: Prague right after OCF (Matthias, Koster, McCool to check the host)
<kaz> 2018-06: Korea hosted by TTA(?)
<kaz> 2018-10: Lyon, France (22-26 Oct 2018 during TPAC)
<kaz> publication planning:
<kaz> Binding Note: A few more cycles (in 6 weeks from now)
<kaz> TD 2nd draft: begin 2018
<kaz> Security Note: first Note next week; updated Note by the end of 2017 and publication in Jan 2018
<kaz> Scripting 2nd draft: need to check with Zoltan
<kaz> PlugFest guideline: end of Jan 2018
<kaz> Architecture 2nd draft: Jan 2018
<kaz> ]]
Kajimoto: Protocol binding and
Security documents, we need to also publish.
... Protocol binding is
informal Note. Michael Koster is not present today,
unfortunately.
... Security Notes, first version is due next week.
Kaz: I am working with Elena on this.
Kajimoto: TD second draft,
beginning of 2018.
... Also, for scripting, we need to check with Zoltan. for the
schedule.
... Architecture 2nd draft, we have 2 important imputs. Inputs
from Mozilla, and other newcomers. "Servient" word is not easy
to understand. This is one.
... Matsukura-san pointed out that, for interoperabiity between
servients, we may need finer description or complementary
documentation. Our experience in Plugfest need to be
imporporated in architecture document.
Kajimoto: I added issues in GitHub in architecture repo.
... Issue# 46.
<kajimoto> https://github.com/w3c/wot-architecture/issues/46
Kajimoto-san is showing slides to explain the issues...
Kajimoto: Terminology
"Servient".
... Two years ago, in 3rd or 4th f2f meeting in Sunnyvale, this
terminology was proposed.
... We use this term as a common word now in this group.
... wrt building block, client and servient are symmetric in
WoT, and can be cascaded.
... Devices are represented as WoT server.
... and also can be cascaded. each can serve as server and
client. So we came up with "Servient" word, back in 2015.
... In WoT IG/WG, we used this term throughout. I understand
this is a novel word for other people.
... In architecture document, "servient", this word needs to be
described in terms of why we introduced this term.
... Servient use cases. Orchiestration and proxy, two use cases
are important.
... WoT client is consuming thing, WoT server provides
thing.
... Turn on AC, put commands in AC. This is one use case.
Panasonic POC proposed, if I say "turn on AC, turn on Hue".
This is one use case.
... We can introduce proxy. This is House in the slide.
... Which can orchestrate all commmands "At home", "Go out",
etc.
... This WoT servient consumes Thing and expose Thing at the
same time. This is first use case.
... Another is proxy use case.
... At plugfest in Osaka at Panasonic, WoT server exposes
thing, and WoT client, both on internet. Panasonic and Fujitsu
both had client and server.
... Fujitsu client could call Panasonic WoT server, and also
the other way around. Siemens implementation was also
used.
... In Dusseldorf, WoT client and servient talked on Internet.
Also, OCF people contributed in Dusseldorf. Smart Things
communicated on Internet.
... But we found WoT client could not access local Thing behind
firewall.
... This became Matsukura-san's motivation.
... NAT traversal technology was used as an example.
... WoT client on internet, using STUN call, through STUN
server, communicate with local server.
... In this case, both need to use STUN server.
... This PPT is about terminology, and also about
Matsukura-san's requirement.
... Second solution introduces "local proxy". This WoT servient
calls STUN server.
... The proxy can call local server. The server does not need
to be customized for STUN.
... With Remote proxy, remote proxy can communicate with STUN
server.
... In this case, client does not need to be customized, but
the server needs to be.
... With Local and remote proxy, proxies use STUN call.
... Local proxy communicates with local server.
... Neither WoT client or WoT server need to be
customized.
... I want to add two use cases, to describe the need for term
"Servient"
... All roles, WoT client, two proxies, WoT server. They are
all "servient".
... I would like to explain the rationale for "servient"
terminology.
... in architecture document.
Kaz: So you are proposing to
continue to use "Servient".
... but We don't have to stick to this term.
... WoT server, WoT client, etc, could be used instead. the
question is what would be easier to understand.
Kajimoto: Architecture document is the basis document on which WoT is built on.
Kaz: Matthias is assigned an
action item related to this topic as well.
... I am not very sure if we need to introduce NAT traversal into the
architecture document itself.
... Core architecture and implementation issues like NAT
traversal should be split.
Kajimoto: Architecture document, we can describe a link to such a document. This approach may be better.
Kaz: We need to check with Matsukura-san.
MLagally: Minutes slide. Node WoT implementation, and reference architecture. I would like to know more on this.
Daniel: We plan to relaunch this in
Eclipse.
... We intend to make it a reference implementation.
... Our goal is to have standard-compliant implementation.
MLagally: Thank you, Daniel.
Kajimoto: By the next F2F, we
should update our public documents.
... I proposed architecture document update plan today.
... script call, TD call, please update document according to
schedule.
... Protocol Template and security need FP WG Notes.
... Please contribute, everyone.
Kaz: Do you plan to update architecture document based on today's presentation, Kajimoto-san?
Kajimoto: Complementary documents, at least I would like to add link to them.
<kaz> kaz: so you're planning to updated the architecture document based on the feedback from the plugfest in burlingame
<kaz> kajimoto: yes
Sebastian: Do you want to add more consideration on proxies in architecture document, right?
Kajimoto: The main idea is try to explain on "servient" terminology.
Sebastian: OK. Thank you.
Kaz: History of servient
terminology, and feedback from Burlingame plugfest, two are
related but separate, which was a bit confusing within the slide today.
... any other topics for today?
(none)
Kaz: In that case, please review the compiled minutes, and SEND IN SLIDES from the Burlingame face-to-face.Kajimoto: Then let's
adjorn.
... the main call will be next Wednesday.
... Thank you.
Daniel: Will we have editor's call today?
Kaz: Let's reconvene next week.
<kaz> [adjourned]