<kaz> scribenick: zkis_
McCool: We have guests who are not
members.
... we also have new members, Farshid from Fraunhofer
<kaz> Apr-6 minutes
Christian: question, discovery should return TDs; is it possible to return TD Templates?
McCool: could be a reasonable extension,
we need to discuss it
... for instance we might want to omit certain information for
privacy reasons
... any objections against accepting the minutes?
no objections, minutes approved
McCool: last week Fraunhofer joined the main call and gave a presentation about a WoT Thing Directory implementation
Fraunhofer's slides on LinkSmart Thing Directory
McCool: there was not much time for questions
Farshid: we could go through questions one by one and discuss
McCool: use case on energy efficiency -> pointing to WoT Architecture use cases
<McCool_> https://github.com/w3c/wot-architecture/tree/master/USE-CASES
<zkis> scribenick: zkis
McCool: we should create an issue to
capture the use case
... looking for possibly related issues
Farshid: the use case is related to energy efficiency of smart buildings
McCool: will create a new issue
<FarshidT> https://github.com/linksmart
new issue on energy efficiency
<McCool_> summary of comments: use case, meta-data, edgeX, ttl, validation with semantic extensions, notification of TD updates, search/filtering mechanism, partial TDs, dids, links for metadata
McCool: any other questions to
discuss?
... prioritizing this, search and filtering should be discussed
Farshid: we evaluate different approaches
for searching
... one candidate is JSON path
... it's on the slides
Fraunhofer's slides on LinkSmart Thing Directory
Farshid: for instance, fetching hrefs of
all TD's
... or search for TDs that have a certain title and fetch only a
few properties
... or search using a regexp
... we could also use comparison operators
McCool: who defines JSON path?
Farshid: it's not a standard
<McCool_> https://goessner.net/articles/JsonPath/
<McCool_> https://github.com/oliveagle/jsonpath
Christian: did you consider others, like SPARQL or JSON queries which are standards?
Farshid: we did a comparison and this was easier for users
McCool: it's been a problem in a past
when we don't rely on standards, as we need to define how we use
it
... we discussed SPARQL and concluded it's too powerful and too
complex
... we used partial TDs for matching
... we could do that with JSON path as well
... would that work?
Farshid: yes it's possible
... also, GraphQL would be possible
Koster: We used searches where we wanted
to know what type annotations there are in the TD
... and this would work great for that
... semantic queries, dissecting TDs from semantic point of
view
McCool: we should define standards for
Thing Directories, and should come up with recommendation for query
language
... we need a set of candidates
... we need to document pros and cons
... at first sight this is appropriate
... we should also check out others
Koster: JSON path is a query language, JSON pointer is different
McCool: anyone knows if there are
standardization plans for JSON path?
... first we need to capture this in an issue
Farshid: do we include protocols in this?
Koster: we need to be able to talk to standard HTTP server using these JSON path or other queries
McCool: yes we should be able to pack it
in a URL
... capturing SPARQL, GraphQL, TD templates, simple keyword search,
JSON path
Koster: SPQRQL and keywork search don't need to know much about the content you're searching in
McCool: we could do keyword search with JSON path
Shreekantha: let's also include XML path
Koster: who are using XPath? examples?
Shreekantha: maybe later
McCool: WoT Scripting supports are few
discovery search options, they are not set in stone
... we need to track how can we support these searches in
Scripting
<McCool_> https://github.com/w3c/wot-discovery/issues/17
McCool: OK, moving to other topics.
... let's postpone the partial TD discussion
McCool: this should be discussed in the TD call
Farshid: OK
McCool: what about optional time to live?
Farshid: we checked this out and optionality was justified
McCool: creating an issue about TTL use cases
Farshid: we expect that a registration created by a human will be managed and removed when it is no longer valid.
McCool: will create the issue and follow up there.
https://github.com/w3c/wot-discovery/issues/18
<McCool_> https://github.com/w3c/wot-discovery/issues/18
McCool: time is up, a lot more things to
discuss, e.g. does Vorto support partial TDs etc
... we will discuss next week
AOB?
[adjourned]