<scribe> scribenick: kaz

Testing

McCool: a few things to discuss
... Ege wants to talk about testing?

Ege: yes
... there is an issue in TD
... updated the assertion tester

McCool: what did you change?

Ege: edited the test for readMultiple

McCool: did you create a GH issue?

<ege> https://github.com/w3c/wot-thing-description/issues/388

Ege: URL of the issue above
... from Princeton. updated 2 days ago

McCool: can regenerate the report
... have not rerun the test for a while
... a lot of recent changes in the TD spec
... @context is now mandatory
... all the TDs need to be updated

Ege: the playground is updated but the test is not

McCool: most TD will fail with that
... assertions to be updated
... let's check what the current spec says

5.2.1 Thing

McCool: @context is mandatory
... the issue is we need tests updated
... and rerun the report
... anything else?

Ege: that's it from me

McCool: besides making @context mandatory
... Time/date fields should be checked

Ege: JSON Schema uses ISO8601 format

McCool: that's good

Demo

McCool: added Intel stuff to the README.md

demo logistics page

Ege: clarification
... we won't have actual demo at TUM?

McCool: right
... made some notes on the README.me

(below the table)

McCool: listed those what I'm sure to bring to the venue
... Devices: 2 lights; 4 LEDs; 1 RGB Led; 2 push buttons; 2 PIR sensors; 2 temp sensors; 1 buzzer
... the scenario should be clear
... bunch of devices listed at the table

Ege: one question about the infrastructure

McCool: we normally set up our own wifi access point
... that's usually set up for a certain network for the demo
... if you have own router, we can configure to use it as well
... we'll have our own local network and wifi connection
... depending on HTTP port 80
... note that port 22 was blocked in Lyon
... should double check with Siemens as well
... 2 issues
... local access
... hostname with .local
... for external access, I'll use tunnel

Ege: hostname.local would be a good solution

McCool: added 2 scenarios
... working at home
... looking at the scenario to charge battery
... the idea is home energy management system
... putting energy to eco-mode, etc.
... also security alert
... PIR sensor, motion sensor, camera, person recognition
... need to figure out if can go out to a different network
... need to do more
... Hitachi also expanded their description
... Ege, if you have multiple demos, you should expand them to multiple rows

Ege: bunch of devices part of a certain ecosystem
... raspberry pis, robots, etc.

McCool: might be better to describe them one by one depending on each scenario
... allocating devices depending on scenario
... OK with allocating another row
... need to find out which devices are used for which scenario
... flash light could be triggered by multiple scenarios
... voice output also could be triggered by multiple scenarios
... all, please keep working on your scenarios/devices
... and document devices for your scenarios on the doc

Kaz: wondering about a bigger scenario including many vendors/devices

McCool: a possibility is the energy management one
... but we're not sure how to do that

Koster: due to various vocabulary?
... fine with having binary mode for eco-mode or not

McCool: yeah
... another question is how to define "eco-mode"
... sometimes the sensor detect "eco-mode" even if there is a person if that person doesn't move
... need some qualification
... we should look into the scenario
... btw, I don't want those scenarios to change TDs

Koster: solving the interoperability problems
... might be some way to define the deferences

McCool: interesting use case

Kaz: any definition within iotschema?

Koster: not yet done
... but something related to economy mode
... I could help you

McCool: currently using some basic binary mode

Koster: pretty straightforward

McCool: anything you can do by the next week?

Koster: can do something

McCool: can look into that then
... somebody use binary mode and somebody may use 1-5, etc.

Koster: related to scale normalization
... kind of like color parameter

McCool: think we should start with boolean pattern
... and then talk about generalization
... should be aware small value would be more eco or not
... for the demo, we should start with something simpler
... anyway, as Kaz mentioned we should think about a big scenario which includes multiple vendors

Koster: that's orchestration

McCool: multiple scenarios could be put on one specific device as well
... the bottom line is showcasing interoperability

Kaz: agree with McCool
... if we could showcase the narrow-waist pattern of the WoT mechanism for interoperability
... that would be really great

McCool: right
... e.g., node-red is a good thing for integrating application logics
... would like Koster to work on the basic definition for eco-mode

Koster: will do

[adjourned]