See also: IRC log
<yingying> scribe: Daniel
<yingying> scribenick: dape
SK: 1. updates for CP w.r.t to
TD
... discussed points are already integrated
... e.g. media-types
... one container for Property, Action, and Event using @type
... local endpoint information can be described also
... e.g, what is Uri, what media-type
<sebastian> http://w3c.github.io/wot/current-practices/wot-practices.html#thing-description
SK: no change for integrating
other contexts.... still possible to use other @types's
... global endpoints are possible also
... idea is to have a global base uri... locally one can define
relative path OR override some information
... e.g., switch from http to coap
DP: for next PlugFest, do we plan to "test" complex type defintitions
SK: lets postpone discussion when
we discuss PlugFest scenarios
... Internal Siemens feedback, especially from Matthias (who
raised issue #287)
... re-use common vocabulary
<yingying> issue 287
SK: endpoints --> link
... uri --> href
TK: Any reason why we have to
follow html example
... endpoints sounds more appropriate
SK: MK just likes to re-use
existing vocabulary
... Dave also mentions that we should try to make TD Web
developer friendly
DP: be cautious about plural form.. link, just singular ?
SK: TK and DP, please
raise/comment issue on github
... be aware that CP document might slightly change the
upcoming days
Naka: href is still in example
3
... will confuse people
SK: Agree
... idea is global vs. local information
... should try to simplify it
... NakaSan, can you raise this as issue?
Naka: OK
SK: Another issue from Alex Owen
#268
... proposed td class field
... references to CSS files
... I am proposing to use @type
... not sure if we really need that new field
... There are some more TD issues... no urgent ones anymore
SK: Dave worked on that
topic
... think we need a more principal discussion
... TD should be independent to programming languages
... Shows slide (WoT TD Basic Assumption)
... Servient1 in JavaScript
... Servient2 in C++
... TD format independent
... JS can wrap JSON
... C++ might map to structs
... focus of TD is to be independent
Uday: strongly support your
argument
... want to be an abstraction layer
... try to be as neutral as possible
DP: Not sure about consequences (JSON vs JSON-LD)
SK: Dave proposes to have TD
similar to Web Developer application layer
... mapping to JS objects
DP: Also would like to mention
that JSON-LD is a set standard and tools are built around
this
... people are used to this tools
SK: I would like to differentiate
between the abstract TD format and the actual applications
working with TDs
... TDs may become complex... people might tend to use tools
for creating TDs
... humans can still edit it... but might not be sensible for
complex scenarios
TK: IF majority of languages is
JS then i would go with JSON
... since we don't know yet a more neutral format is better
SK: Dave will present his idea in the afternoon call
Uday: w.r.t. to organization .... how many further TD calls
SK: I expect one more
... Let's try to resolve issues
... hope we can freeze document next week
... not sure about API changes
... we will then tag and publish a freezed CP document