See also: IRC log
sebastian: thing description
needs to deal with contrained devices e.g low bandwidth, low
memory
... description needs to be minimal description of
capability
... thing description needs to be on the device itself
... displays TD mdoel developed 3 months ago in california
<jcdufourd> scribe: kerry
sebastian: play around with this
and do some implementations
... plugfest tomorrow in WoT meeting
... do we need more description?
... TD contains metadata ,three kinds of interaction (property,
action, events)
<jcdufourd> before rrsagent was invited, kerry introduced spatial data on the web http://www.w3.org/2015/spatial/wiki/Main_Page
sebastian: metadata is used to
provide some generic information which may not be relevant at
runtime e.g. for setting up communication
... property actions and events are used at run time
e.g. temperature value name, output data type, writeable true/.false
scribe: stability (e.g. each 5
mins)
... ACtions invoke actions which might take a while ===
actuation
... using json-ld for representing data
... show exampkle for a tradffic light
... includes list of protocols, ranked, encodings (e.g. json),
turn green, turn red, actions
... how to bind thing description to a protocol?
... plufest show that thing description is easy to interact
with, without implementation
... how to enrich with further context?
... how to show the meaning of more properties...
... have more semantics in next plugfest?
... e.g. location could go in metadata...
... very constrained to work on small devices
... some experience eith ssn that it is too big (30 triples for
one property)...
... maybr just somje subsets...
... data is missing...
DarkoAnicic: need best practices for scale up from the minimal to more context
armin: do you have a memory limit for metadata? that could be used to say describe temp sensor?
DarkoAnicic: difficult -- that depends what derialisation comes back with
akeranen: 80 b or multiple
thereof
... 10 KiB ram 100kb flash (internet class 1)
DarkoAnicic: we son't necessarily
need to use ssn on the device --- just an additional
layer
... so it does not have to be ssn on the device -- only the
TD
Sebastian_Kaebisch: goal is to
have as much as possible on the device but only what is really
really required, including protocool and encoding
... just first working assumption
... exi and binary json-ld will ber important
... only a few bytes
DarkoAnicic: offer this
extraction capability as an approach then
... examples of how this might be used
... ssn needs protocols and "wrtiable" and "encoding" etc
Sebastian_Kaebisch: may need a standard for encoding (discussion) using json-ld
deb: ask to finsd out protocol before you know what the protocol is?
Sebastian_Kaebisch: can use a uri
to request it from (or another way) not yet resilved, but
relying an a bootstrapping defualt
... eg always start with http to get started
akeranen: binary json is?
Sebastian_Kaebisch: exi will also
be supporting json
... exi will provide support for schema... semantics have a lot
of strings and exi does this well (seborg(?) does not)
... if you want a simple encoder complexity is same --- getting
the data is the priority
... lets have a look ath this approach to extract from ssn
armin: we could also play with identifiers to reduce the space
DarkoAnicic: may need to redesign ssn and then build up
akeranen: discovery.... one standard mechanisim for doing discovery
DarkoAnicic: interest group needs
tech landscape suvey to provide to WG
... plugfest attndance -- next plugfest meting in January
<akeranen> The standard mechanism: https://tools.ietf.org/html/rfc6690#section-4
<akeranen> CoRE weblinks and .well-known/core
summary: keep in conact. SDW can accommodate wot requirements for ssn. Key requirement is minimal space simplicity
This is scribe.perl Revision: 1.140 of Date: 2014-11-06 18:16:30 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/woth/with/ Succeeded: s/80 kb/80 b/ Succeeded: s/20kb ram/10 KiB ram/ Found Scribe: kerry Inferring ScribeNick: kerry WARNING: No "Topic:" lines found. Present: DarkoAnicic Sebastian_Kaebisch akeranen armin jcdufourd oonishi sebastian taisuke Debbie_Dahl Got date from IRC log name: 28 Oct 2015 Guessing minutes URL: http://www.w3.org/2015/10/28-swot-minutes.html People with action items: WARNING: No "Topic: ..." lines found! Resulting HTML may have an empty (invalid) <ol>...</ol>. Explanation: "Topic: ..." lines are used to indicate the start of new discussion topics or agenda items, such as: <dbooth> Topic: Review of Amy's report[End of scribe.perl diagnostic output]