Meeting minutes
Agenda
<kaz> Agenda for today
1. review PRs and merge as appropriate
2. Implementation reports
3. How to test profiles?
mmc: Want to add all new test data under events/Online ...09
… will copy over input data to In putData. If you need name changes, please change in both places.
PR review
PR 374
<kaz> PR 374 - TUM manual.csv
mmc: that's an old PR, still WIP
ege: will check with Fady
<mlagally> s/checkwith/check with/
PR #416
<kaz> PR 416 - adding TD with a problem details response
ml: This is a TD with additional response for problem details
mmc: this is a testcase for TD and for Profile
… we also need this under "Profile"
ml: Will also create and merge a PR which puts it under Profile
PR #417
<kaz> PR 417 - adding TD with async action
<kaz> (discussion on the type of "status")
< discussion between Ege, ML, MMC about validation>
<kaz> mmc: (adds suggestions to WoTWebThing+problemDetails+actionStatus.td.jsonld within PR 417)
mmc: We need a JSON Schema which does strict checks
ege: Jsonschema is also permissive - you can try with correct and incorrect schemas
mmc: we should make our tooling stricter
ege: I can write a small script for the TD to generate a strict version
ege: we don't have all features of JSON Schema in a TD
(discussion on href with anyURL)
kaz: for the technical discussion, JSON schema could be used as part of the necessary validation, if it is useful and usable
… however, if we want to really validate the inputted TD strictly, we need a better mechanism
mmc: kaz: second point, please use q+ on IRC to manage the speakers :) is not very powerful, we talked about other means for tests
… don't completely trust the validator
… the PR is a useful testcase - let's keep it open
PR 409
<kaz> PR 409 - Change procedure to include Consumers
mmc: This is a suggestion to have results for consumers
… This would be useful for TDs as well
… I have already split up implementations for things and consumers
ege: I mentioned we should test it during the last testing call
… I was asked to create a PR in the last testing call
mmc: it would be much cleaner if we had two CSV files
… do we have time to do it
… It may need a bit more thought to split up the assertions
ml: splitting the assertion into two would be too much
… we're now talking about Consumers
mmc: currently everything in Profile is only manually testes
ml: The TD has assertions on Consumers, how were these tested?
ege: This is tested in the TD via manual assertion
… we have neither done a plugfest on Profiles, the Profile has constraints on the consumer
… we have to test interop of the spec.
… how do we want to test Profiles?
… I could just provide manual.csv, but how to be sure?
mmc: So far we trust submitters of manual.csv, there's always more testing that could be done
… I don't want to add additional testing given the time constraints
… there's some discussion to defer REC publication of the Profile
… We may consider the "Profile" feature of the TD at risk
mmc: Let's keep the PR open and rediscuss next time
PR 374 - revisited
mmc: Fady, could you update the TUM manual.csv for the Testfest 09?
fady: will do
Implementation Report
Architecture
wot-architecture PR 849 - Oct 2022 Testfest Updates to Implementation Report
mmc: I have a PR #849, will update it in the description as we go.
… It has a link to a rendered version
… we have 13 implementation descriptions, some need some work
… will poke people to fill some blanks
… we got some new results for Architecture - We still could get 8 more results
… Some responses should be checked again, e.g. in the security considerations for use of encryption in public networks
… Please take a look offline and review your responses
<McCool> https://
<McCool> Draft Implementation for Architecture by Statically
TD
<kaz> wot-thing-description PR 1706 - Oct 2022 Testfest Updates to Implementation Report
mmc: I also did the TD, there's a couple of new things
… around JSON pointers
… I18n is a bit tricky -
<Zakim> Ege, you wanted to I have asked Ditto implementation to provide the manual.csv for architecture
ege: You don't have to ping ditto, I already talked to them
<kaz> @@@move Ege's comments before TD implementation report
mmc: (Reviews details of the implementation report)
mmc: Thing model seems to be missing input
ege: perhaps an old version?
mmc: Will run the scripts today, let's see if the problem goes away
cris: tm-versioning-1 and tm-versioning-2 should be manual
<McCool> Draft Implementation for TD by Statically
Discovery
mmc: please review PR and comment on it
<kaz> [adjourned]