Meeting minutes
Previous minutes
Minutes approved
PRs
Fix #355: change use cases,
PR 375 - Fix #355: change use cases
McCool: the text "The following scripting use case scenarios are supported" is fine
Kaz: agree. so would suggest the title also should say "Scripting Use Case Scenarios".
Daniel: suggest to wait for Zoltans feedback
Issues
Using a callback-based approach for Discovery?,
364 - Using a callback-based approach for Discovery?
McCool: discovery API should deal with all possible options
<kaz> 10.2 The DiscoveryMethod enumeration
Daniel: we currently have direct and directory
Cristiano: w.r.t. registration is left to runtime
… shall we make this explicit
McCool: self- description is another argument
… like .well-known
McCool: need to re-read the scripting spec
… generate issues for missing features
Cristiano: I suggest to wait for feature freeze of discovery
McCool: should happen soon
Cristiano: exploration methods only ?
McCool: simple call, just "discover"
Cristiano: problematic to be implemented
McCool: testing of discovery
… client-side implementations
… "introduction" is open-ended
… DNSD should be supported
… need external management API
Daniel: Separate API
Cristiano: filtering TDs (query) ?
… tricky to filter
McCool: Filters are a real problem
… SPARQL is heavy weight
… JSONPath is not ready yet
… XPath lacks implementations
… we ended up not having mandatory query mechanism
… should support filter like JSON-Path in node-wot
… do apply filters locally
Cristiano: not suggest JSONPath on client-side
… implementation burden
Cristiano: application knows TD
… how can I fetch TD
… using discovery API?
McCool: optional argument is fine with URL
… description of discovery process is missing
… set of TDs
Cristiano: discovery process would be great, we just follow
… Scripting task-force needs to review the process
McCool: need to still describe it
… proposed text in introduction section
Cristiano: Sounds great
Daniel: different discovery entry points
McCool: options to set
… we might want to have multiple options
Cristiano: JS allows to check for which method is supported
McCool: configuration option might be better
McCool: "automatic" method is missing
… I am working on integrating discovery services
Daniel: unsure about "automatic" method
McCool: question about limiting the scope
… in local networks it might be fine
we lack geo-spatial control
McCool: discovery does support mDNS etc
… "local" could be local discovery (like "automatic" in local network)
Cristiano: what about BT ?
McCool: is local
Cristiano: What about process?
… partially describe it
McCool: there is an informative description
… should make it more formal
… but we are running out-of-time
Daniel: timeline?
McCool: overdue, last week
… best we can do, having informative description
<McCool needs to leave>
McCool: will create issue with adding "local" back
… will re-read the document
… self-description
Cristiano: left to implementation?
McCool: when exposing ... like under .well-known
… how to decide where to broadcast
Cristiano: opposition in scripting: expose() makes it available
McCool: self description should not be required.. there might be cases you don't want
Daniel: Issue with having *one* TD only on .well-known
McCool: working on fixing it
… look for simple mechanims
… thing-links
… pointing to other TDs ... index on host
… several runtimes on same host is problematic
Cristiano: Suggest to wait for feature freeze
… w.r.t. "local" ... I am surprised .. but god the point from McCool
… should re-iterate on API... e.g. from Jan
Cristiano: should try to have another call with MMC
<kaz> [adjourned]