<kaz> scribenick: McCool
https://www.w3.org/2020/07/15-wot-td-minutes.html
https://www.w3.org/2020/07/22-wot-td-minutes.html
no objections to publish July 15 minutes
no objections to publish July 22 minutes
sebastian not available 19/8, 26/8, and 3/9
Sebastian: will Taki be available?
Taki: yes, I will be available
Sebastian: it would be ok to have a break, too
Taki: in that case, let's cancel the 19th and maybe the 26th
Sebastian: definitely cancel 19; taki
can decide and announce in the main call
... based on open issues
McCool: I will also record cancellations in main wiki page
https://www.w3.org/WoT/IG/wiki/Main_WoT_WebConf#Cancellations
for TD1.1 https://github.com/w3c/wot-thing-description/labels/FPWD%201.1
json schema issue has been fixed - https://github.com/w3c/wot-thing-description/pull/896
<Ege> brb in 2 mins
Sebastian: have asked Daniel to fix
render script
... to generate based on ontology files
McCool: other PRs also need to know how render works, eg in the ontology file or in the template
Sebastian: we also want the rendered diagrams based on the ontology
Cristiano: we also talked about
contentType, etc.
... where will these changes go?
Sebastian: actually I forgot where this was supposed to go; I will update the issue and deal with this as well in the PR
<sebastian> https://github.com/w3c/wot-thing-description/issues/912
look at oauth2 pr
https://github.com/w3c/wot-thing-description/pull/927
issue with using an ontology (formal vocab) for flow names
<trackbot> Sorry, but no Tracker is associated with this channel.
would want to do this in a backward compatible way, and allow external vocab to define extensions
eg it needs to be an open enumeration, not closed
Sebastian: let's discuss under the PR
next, issue 937
https://github.com/w3c/wot-thing-description/pull/937
seems there is an issue in the context files
Sebastian: some redundancy
... in particular, an @id under security seems wrong
McCool: certainly I may have made an error here
Sebastian: ok, I will check it
... also moving TDT to main section https://github.com/w3c/wot-thing-description/pull/938
... these are all the PRs that would go into TD 1.1
McCool: so this is just a move, not a
redefinition?
... I do think we should loosen the definition a little
... for example, it should not be mandatory that security is
omitted
... maybe we should allow URI templates in forms, etc
how to version scripting API separately from TD versioning
Cristiano: also TD and scripting
versions are coupled at present
... need to know which version of TD a version of the scripting
API works with
... how do we deal with "semantic" versioning when we also have
"dated" URIs for context files?
Sebastian: kaz, do you have any experience with how this is handled?
Kaz: we can look into JS and
HTML
... but the assumption there is that everything is backward
compatible
Daniel: example is XML schema,
version 1.1 still uses the same namespace
... so looking at the namespace can't be used to determine the
version
McCool: I personally would like to
see a scheme like .../v1 referring to latest v1, v1/0 and v1/1
referring to specific versions
... unfortunately the date in the URL messes that scheme up
Sebastian: we should ask the JSON-LD group about this topic
<kaz> <script type="application/javascript;version=x.x">
Kaz: if needed we can use something like the HTML version tag, e.g. using extended context, or adding specific tag
McCool: assume we are focusing on TD versioning for now
Kaz: good JSON-LD joint meeting topic
Sebastian: agree
McCool: we should reach out to them and schedule that
Sebastian: thanks ege for moving issues
over from the old wot repo
... there are some interesting ones
https://github.com/w3c/wot-thing-description/issues/936 using links for UIs, IDE
Sebastian: also relates to Schema,
but we have also introduced title and descriptiong
... but an open question is how to provide icons, etc.
... cannot be solved by linking, since links are only at top
level, not at the interaction level
... maybe it can be dealt with using semantic annotations
Sebastian: please share ideas on issue
Ege: I also think semantic annotations would be the easiest way to do this
<Ege> https://schema.org/logo
McCool: I think for "generic" icons
associating it with a semantic tag might be feasible, but for
branded icons it is tricker
... logo thing seems more like an extension than an
annotation
... if it's a common extension, maybe it should be in the main
spec
... in HTML it makes sense as an annotation since you can
annotate an existing image tag with microdata
issue https://github.com/w3c/wot-thing-description/issues/935
McCool: class field to be able to modify several properties at once
Sebastian: we may also be handling this in different ways now
McCool: I think this is more of a
scripting API issue if it is about manipulating multiple Things
at once
... eg the result of a query on a particular annotation
<dape> https://github.com/w3c/wot-scripting-api/issues/204
Daniel: this does not seem like it should be in the core spec of scripting though
McCool: I agree, it can be
implemented using other features we are already working on, in
particular semantic annotations and directories queries that
return a set of results
... but we might have to consider some tweaks to make it
possible to submit queries to directories without revealing the
URL of directories (for privacy reasons, eg. to avoid
fingerprinting...)
<inserted> related issue on wot repo
Sebastian: my reading is that this
example is addressed with our current spec
... except for global definitions of types
... example uses SSE defined in a "protocol" field
... for example; but this specific case we have subprotocols
now
<sebastian> https://github.com/w3c/wot-thing-description/issues/307
Sebastian: then there is the use of
$ref etc for definitions
... there are points on our roadmap that will address the
compound values
<inserted> https://github.com/w3c/wot-thing-description/issues/934
Daniel: but not sure about streams; there is another open issue on streams I believe
Sebastian: let's keep it open, and discuss
Ege: looked into in detail
<sebastian> https://github.com/w3c/wot-binding-templates/pull/98
Ege: is missing concept of
default modbus binding
... e.g. is read property reading a bit a word?
... also, do vocabulary files need an "editor" at the
top?
... respec gives me a warning if there is no editor
Cristiano: also the examples still need to be updated; the uppercase values need to be camel-case
Sebastian: would check that the terms use correspond to normal modbus terms
Cristiano: may also have to update node-wot implementations, may not have used the exact same terms
Kaz: I thought there are two
resources here, (1) machine-readable vocabulary definition and
(2) human-readable document describes that. If we want to
publish the human-readable document as a W3C Note, then we DO
need an editor
... applies to human-readable documents
Cristiano: about length and byte length, for node-wot, length is defined in content type
McCool: I am also wondering about extensions to data schema for bit length
Ege: better to distinguish data schema from protocol information
Cristiano: problem is that if returns
something like json, then it is redundant to have the
length
... but if the content type is an octet stream, we should
include the length in the content type
Sebastian: seems like it needs more
deliberation; have taken notes in the issue
... we will watch for his replies
... author only worked loosely with Ege
Kaz: maybe this is overkill, but
do we want to worry about "livestreams" where the end is not
known?
... possibly related to media discussion
Cristiano: not applicable to
MODBUS
... might apply to other protocols however, e.g. media
McCool: assume in the case of
livestream we could use a content type without a length
... ends when socket is closed
Sebastian: ok, done, move to adjourn
<kaz> [adjourned]