Meeting minutes
previous minutes
<kaz> July-6
Ege: can you update the initials?
Kaz: will do
Ege: now we have the pre-commit hook to render the spec automatically
… we also discuss about how to return the payload in a writeproperty
… discussion was deferred
… same as initial connection
… time series and historical data is now a topic of the next charter
… operation pairs was deferred too
… minor edits in the binding templates
… then we finished
… I don't see any issue with the minutes
Kaz: name fixed
Ege: they look good
… any objection to publish?
daniel: noticed that Sebastian name is misspelled
Kaz: fixed
Ege: alright minutes approved
Agenda
Ege: any topic to add?
… ok none
publication status
document check status
Ege: there is a PR from @kaz
… any updates
<kaz> PR 1573 - 4-wd publication draft based on the Editor's draft as of 31 May 2022
Kaz: it is already updated
… daniel created another PR to remove the extra CR directory
Ege: but the publication is not yet on w3c servers
Kaz: there are some editorial changes to do
Ege: ok thank you
daniel: kaz did you integrated the changes for fixing some links and references?
removing the obsolete 4-cr directory
Kaz: please merge Daniel's PR first
<kaz> PR 1569 - remove out-dated folder publication/ver11/4-cr
Ege: so it is a minor pr
… it is removing 4-cr
… merging any objections?
… merged
respec errors
Ege: in 4-wd I'm seeing some errors that should be fixed
daniel: yes one was alredy fixed by Michael and the other fix is in flight in one of my PRs
Ege: what about the red border of W3C ?
Kaz: it is just CSS once we move to static css it will be removed
daniel: I used the respec function to create the static CSS
Ege: ok we can proceed
Kaz: I'll look into the detail with the static HTML
Pull Requests
PR 1574
<kaz> PR 1574 - Update Shacl
Ege: supersede other PRs
… any objections to merge it?
… ok merged
PR 1568
<kaz> PR 1568 - refactor: explicitly state unicode code points and give example
Ege: approved from w3c team
… it specifies better the usage of unicode
… any objections?
… ok merged
PR 1570
<kaz> PR 1570 - fix: add link to form
Ege: Daniel is fixing those errors that we saw in 4-WD
… any objections?
… alright merged
PR 1571
<kaz> PR 1571 - fix tm:required example
Ege: examples had wrong json pointers
… some json schema validators accepted JSONPointers with #
… probably is because of RFC which is not clear
jan: there are implications also for tm:ref
… it assumes that you can use #
Ege: but the text on tm:ref was a little bit different
jan: but my assumption was that you can omit the URI
Ege: it is a little bit inconsistent too
… we can fix it by changing the assertion of tm:required
… we are somewhat align with json schema
jan: the approach that json schema uses it seems fine
… about the inconsistency I'm not sure it creates some problem
Ege: it might cause problems in the converter
jan: it should be an easy check
Ege: I'd open an issue about the inconsistency
… later on
… should we merge it?
… merge it
PR 1564
<kaz> PR 1564 - explain contentType usage
Ege: I tried to explain how to use content type inside a TD
… the discussion started in protocol binding templates
… I've created a table
… it is a little too verbose and probably hard to understad from newcomers
Cristiano: why don't add examples about the difference cases?
Ege: Will do
… I will merge it
daniel: what about the matrix idea that you had
Ege: I can do it as an alternative
PR 1575
<kaz> PR 1575 - Remove assertive language from TD Processor definition
Ege: the TD Processor contained assertions
… with also assertive text forcing to use validation
… it was not correct since validation is a optional step in TD
PR 1576
<kaz> PR 1576 - Fix local biblio references
Ege: I checked our local biblio
… there were some problems with the references
… architecture ref used the old author list
… wot security guidelines
… should it be updated?
McCool: no immediate plans to update it
… can we use github.io ref?
Ege: we can't it should be static url
… meanwhile I've just updated the date
McCool: you can link to the blob
… but it will not be rendered
… still I would use the github.io reference
Kaz: regarding updating the security guidelines, do we really have any problems if the reference to that document is not really updated ?
Ege: good point
Kaz: general considerations are contained in wot-security-guidelines
… but normative assertions should be inside the TD document
Ege: exactly
Kaz: therefore it is an informative link
… not critical
Ege: ok
… there were also format issues
… any objections ?
… ok!
… merged
PR 1577
<kaz> PR 1577 - fix contentType of TM links
Ege: content-type in tm:extends examples was wrong
… the PR fixes this
PR 1578
<kaz> PR 1578 - Add text about tm respecting td information model
Ege: it wasn't clear if TD types apply also to TMs
… the PR adds a small text explaining the correct behavior
issues
issue 1548
<kaz> Issue 1548 - Oracle review comments on CR draft - Part 1
Ege: review from Lagally
… I've commented
… with objections to his points
… for example we can use our own media types
+1
McCool: I should keep an eye on RFC assertions inside tables
… it create a sub assertion that it tricky to manage
… it is always better to put the assertion out from the table
Ege: we don't say anything about empty forms
… we can combine different operations
… too late for add the cancelAllActions
… events might have empty payloads
Cristiano: : +1
Ege: regarding versionInfo we should recommend to use semVersioning
… if that is the case we should explain what PATCH, MINOR, MAJOR means in TD development
Cristiano: reading the text it seems that we already recommend it
Kaz: we don't have Michael Lagally here today, I suggest to postpone the discussion for the next time
… and invite him
Issue 1400
Ege: we agreed on Thomas points
… and it is quite the same in json schema
… we decided to not doing anything
Cristiano: are you sure? I recall some action item
Ege: right, we need to probably discuss about tm:required default value
… creating an issue
… is there any additional topic about TD?
McCool: yes there is a PR about the new implementation report
Ege: ok before doing that
issue 1294
<kaz> Issue 1294 - Default value section misses op default for Thing
Ege: form without op in the root level are useless
… therefore I create a PR
… to add this assertion
… it is a normative change, not sure it is ok
McCool: it is adding a new assertion,
… it is critical?
Ege: should not break testing
McCool: saying it must be there is an useful validation check
McCool: not a big deal
Cristiano: was json schema updated?
Ege: yes
implementation report
<kaz> Preview for the Implementation Report
McCool: I've updated node-wot
… I fixed a bug with reports of Siemens Logilab
… reports
… philips-hue stuff is there too
… things are pretty good
… first three should be moved to manual.csv
… bearer tokens are zero
Ege: did we have it
… in the previous version?
McCool: it was an experimental thing
Cristiano: we are using extensions for bearer tokens in node-wot
McCool: ok
McCool: combo schemes should be higher than 0
… probably a validation problem
Ege: right
McCool: device flow might be still at risk
McCool: contentCoding assertion is about compression
… should be easy to implement
Ege: about default Daniel did something
daniel: yeah
McCool: in field is failling
Ege: maybe nobody used digest
McCool: I did one
… node-wot does not support it
McCool: one of
… I'm surprised
… there might be an error
Ege: right
Ege: regarding cancelaction did you include oracle implementation ?
McCool: let me check
… I'll check offline
Ege: don't node-wot support unobserveAll ?
Cristiano: I think so
Ege: it seems that security is an outliner
McCool: right
… other notable mention is schemaDefinitions
… it should be defined in all the directories
Ege: about icon size
… it easy we can add thingweb logo
… daniel can you add it?
McCool: remember that we have the todo.csv
… we need to keep it updated
… notice that the assertion id is tm-extend but the vocabulary says tm:extends
McCool: on the manual assertions
… a lot of these are pretty easy
… for example fetching context file it easy to do
… but I don't know if any of the implementations does so
Ege: the new ones need another pass
McCool: right
… I suggest mergining this now
… and do updates later
… any objections ?
Kaz: updated report is good, but we should think about which part of work is handled by TD TF, and which by testing TF
McCool: still need to look at testing tools in each deliverable's task
McCool: suggest create issues as appropriate in wot-testing repo
Kaz: it would be great if we could summarize what we've done so that people can use it as a best practice.
Ege: (merged implementation report PR)
Binding Templates
PR 156 and 158
<kaz> PR 156 - Improve the mqtt protocol binding template
Cristiano: improvements to MQTT binding
<kaz> PR 158 - Improve CoAP document
Ege: also PR 158
… PR 158 has gotten a lot of feedback and was improved
… number of small changes
… Jan and Philipp provided a lot of feedback
… however, major issue still is normative content in examples
klaus: if this is the only outstanding problem, propose we merge this PR and create an issue for this specific problem
Ege: also, note that we CAN use normative language in these documents, even though it is a NOTE
… but for CoAP and Content-Format in particular, note that IANA registration is necessary
… can't just have arbitrary strings, need to use an assigned number
… also point about default value; what exactly is meant here?
Jan: not sure, whether this should have a default value
klaus: if you don't put content type in the form, TD has a default value
… also request payload and response payload
Jan: ok, let's resolve the thread
Ege: thanks; will merge 158
Ege: and related issue was automatically closed
PR 156 - revisited
<kaz> PR 156 - Improve the mqtt protocol binding template
Ege: four commits
… look at recent changes
Ege: main changes was localbiblio, move refs from abstract to intro
… uri scheme part in combination with topics
Cristiano: should use topic name, filter, and form
… and provided an example to explain better
Cristiano: think it is an improvement, can do better in the future
… also issue with parts generated from ontology
… EXAMPL 4, complex property
McCool: want to note also that "Web of Things (WoT)" needs to be added to title for all binding documents...
Ege: good point, let's make an issue, deal with this in another PR
Other issues
Ege: note need to fix Koster's affiliation
PR 156 - final check and merge
Ege: think we are done with this, will merge, create an issue about changing the titles
<kaz> (merged)
AOB
Cristiano: maybe add an issue about references in MQTT binding
… in MQTT vocabulary
… respec does not say anything, but needs to be fixed
Ege: will create issue to fix link errors in MQTT binding doc
… issue 166 created
<kaz> [adjourned]