IRC log of wot-td on 2017-04-28

Timestamps are in UTC.

06:36:59 [RRSAgent]
RRSAgent has joined #wot-td
06:36:59 [RRSAgent]
logging to
06:37:15 [MariaPoveda]
MariaPoveda has joined #wot-td
06:38:28 [kaz]
scribe: Daniel
06:38:30 [dape]
scribe: dape
06:38:33 [kaz]
scribenick: dape
06:39:02 [dape]
SK: goes over agenda
06:39:06 [dape]
.. action items
06:39:13 [dape]
... break out sessions in Osaka
06:39:29 [kaz]
present+ Kaz_Ashimura, Sebastian_Kaebisch, Daniel_Peintner, Feng_Zhang, Maria_Poveda, Yongjing_Zhang, Zoltan_Kis
06:39:30 [dape]
... TD model discussion (report from call this week)
06:39:55 [dape]
... Zoltan joined call.. overview about management discussion
06:40:03 [dape]
... any other agenda point?
06:40:14 [dape]
TOPIC: Action Items
06:40:34 [dape]
SK: TD and semantic annotations?
06:41:01 [ohura]
ohura has joined #wot-td
06:42:23 [dape]
DP: started issue on JSON schema
06:42:39 [dape]
... Victor plans to add example to make our point even clearer
06:43:08 [dape]
SK: plan to have more concrete proposal for Osaka
06:43:21 [dape]
SK: next AI is about namespaces in TD
06:43:32 [MariaPoveda]
there is an issue about this
06:43:41 [sebastian]
next issue
06:44:36 [dape]
SK: Confusion whether we need year and month in namespace
06:44:51 [dape]
... it seems the reserved name is fine
06:44:55 [dape]
06:45:31 [dape]
SK:I am missing the discussion which vocabulary goes in which NS
06:45:39 [kaz]
present+ Achille_Zappa
06:45:55 [dape]
... we need clarification with that regard
06:46:00 [kaz]
present+ Fernando_Serena, Victor_Charpenay
06:46:12 [dape]
Maria: use Excel file so far..
06:46:26 [dape]
... it seems that WoT NS is the only one necessary
06:46:31 [dape]
... rest can be out of WoT TD
06:46:40 [MariaPoveda]
06:47:00 [dape]
... will point to this spreadsheet in github
06:47:46 [dape]
SK: another possibility would be to create PR by for example creating subfolder with examples etc cetera
06:48:55 [dape]
DP: question about evolution of TD vocabulary.. no mean to indicate change with year/month
06:49:31 [dape]
Maria: what I have seen that main NS points to latest version
06:49:47 [dape]
... one with version and one without year/month
06:50:10 [dape]
SK: Might be an issue for already deployed devices
06:50:11 [DarkoAnicic]
DarkoAnicic has joined #wot-td
06:50:29 [DarkoAnicic]
present+ DarkoAnicic
06:50:56 [kaz]
06:51:04 [dape]
Maria: Let's look at what spatial data people are doing
06:51:41 [dape]
Fernando: if one need specific version.. need to add month/year... of not use canonical version
06:51:55 [achille_z]
achille_z has joined #wot-td
06:53:02 [dape]
06:53:05 [dape]
ack dape
06:53:17 [zkis]
zkis has joined #wot-td
06:53:57 [dape]
Fernando: will add my comments to github issue
06:54:01 [kaz]
06:54:21 [dape]
Kaz: W3C has guidelines
06:55:39 [kaz]
06:55:40 [dape]
... recommendation is YYYY/MM ... /ns etc is fine also
06:56:41 [dape]
... will check within W3C team also
06:56:51 [dape]
... should clarify our requirements first
06:58:07 [achille_z]
present+ Achille Zappa
06:58:35 [dape]
SK: Fernando suggests to have both... canonical uri points to latest YYYY/MM uri
06:58:54 [kaz]
06:59:08 [dape]
Kaz: namespace uri and URL might be different... need to check though
06:59:54 [dape]
SK: please use issue to share your view
07:00:09 [dape]
TOPIC: breakout sessions in Osaka
07:00:12 [dape]
07:00:16 [dape]
ach kaz
07:00:19 [dape]
ack kaz
07:00:32 [dape]
SK: I initiated 3 topics
07:00:39 [dape],_May_2017,_Osaka,_Japan#Plenary.2FBreakout_Topics
07:00:50 [Fernando]
Fernando has joined #wot-td
07:00:56 [kaz]
s/namespace uri and URL might be different.../we don't usually configure a link from canonical namespace uri to dated namespace uri/
07:00:59 [dape]
SK: 1. TD model reflection (What is missing? How about the security and lifecycle aspect?)
07:01:51 [dape]
... we have to think about security and life cycle
07:02:11 [dape]
SK: 1. Data type with semantic annotation based on JSON Schema
07:02:15 [kaz]
s/need to check though/need to check, though. btw, maybe using "" as the namespace URI would make sense because we use "wot-thing-description" for the TD GitHub repository
07:02:24 [dape]
s/1. Data type with semantic annotation based on JSON Schema/2. Data type with semantic annotation based on JSON Schema
07:03:00 [dape]
... would like to have a concrete proposal till Osaka
07:03:06 [kaz]
rrsagent, make log public
07:03:10 [kaz]
rrsagent, draft minute
07:03:10 [RRSAgent]
I'm logging. I don't understand 'draft minute', kaz. Try /msg RRSAgent help
07:03:11 [kaz]
rrsagent, draft minutes
07:03:11 [RRSAgent]
I have made the request to generate kaz
07:03:18 [dape]
sK: 3. start discussion about TD serializations
07:03:22 [kaz]
s/rrsagent, draft minute//
07:03:40 [dape]
... collect ideas about what serialization...requirements etc cetera
07:03:54 [DarkoAnicic]
07:04:15 [dape]
SK: please add comments/topics
07:04:23 [kaz]
07:04:25 [kaz]
07:04:29 [kaz]
ack D
07:04:56 [kaz]
Chair: Sebastian
07:05:04 [dape]
Darko: Victor and myself discussed SSN ontology... which will become standard soon..
07:05:20 [dape]
... will be Breakout... also as part of TD breakout
07:05:39 [dape]
... maybe joined breakout?
07:06:13 [dape]
SK: To show it can e aligned? Sounds fine!
07:06:28 [dape]
TOPIC: TD model discussion (cont.)
07:06:40 [dape]
SK: good discussion on Wednesday
07:07:17 [dape]
07:08:00 [dape]
outcome see figure..
07:08:18 [dape]
Maria: Pull request still needed..
07:08:31 [dape]
s/Pull request/merging pull request
07:08:56 [dape]
SK: model based on Marias proposal
07:09:01 [Victor]
Victor has joined #wot-td
07:09:27 [dape]
... top class Thing: renaming took place
07:11:10 [dape]
... in general we removed "longer" names
07:12:13 [dape]
... also made clear that uri's MUST be of class URI etc.. not just a string
07:13:05 [dape]
... discussed also whether zero interactions make sense
07:13:35 [dape]
... seems the case given that there might be a simple "overall" TD with just base information without actual interactions
07:13:56 [dape]
... similar renaming took place in other places
07:14:36 [kaz]
rrsagent, draft minutes
07:14:36 [RRSAgent]
I have made the request to generate kaz
07:15:04 [dape]
... also added placeholders (optional elements) ... e.g., communication protocol
07:15:39 [dape]
... other vocabularies can be added
07:16:24 [dape]
... "blue" in figure means it gets removed
07:16:44 [dape]
... "red" stays for modifications
07:16:55 [MariaPoveda]
i still disagree about some renamings
07:17:22 [dape]
Fernando: still unsure about renaming
07:18:04 [dape]
... semantic and schema/syntax names should not be coupled
07:18:28 [dape]
... ontology should not be aligned with schema
07:19:07 [dape]
SK: developers might have different background
07:19:34 [Victor]
07:19:44 [dape]
... same naming facilitates the understanding
07:20:23 [dape]
Maria: we can provide mappings
07:21:41 [MariaPoveda]
The more information are provided in the schema the easier for people from different backgrounds would be to understand the model, simplilfiying names get exactly the opposite, it is more difficult to understand de directionability of the properties
07:21:42 [dape]
Fernando: the approach is to simple... people might want to extend model and use different naming
07:23:03 [dape]
SK: fear additional complexity for these bindings
07:23:32 [kaz_]
kaz_ has joined #wot-td
07:24:47 [dape]
Maria: alignment can be done in 2 ways: align names or specify binding
07:25:04 [dape]
... not sure why one is better than the other
07:25:59 [dape]
.... coupling might be good for WoT peoples.. nevertheless some other people will come up with new syntaxes anyway... and we need mappings anyway
07:26:19 [dape]
07:27:22 [dape]
Fernando: for us the main consumer will be machines.. not humans...
07:27:41 [dape]
... due to the binding this works fine
07:27:49 [dape]
.... we try to avoid strong coupling
07:28:16 [Victor]
ah ok...
07:29:15 [dape]
SK: I agree semantic description not meant for humans..
07:29:25 [dape]
... intent for TD is to have actually both
07:29:38 [Victor]
right. In fact, what you described is what has been done sor far. The JSON-LD context maps verbose OWL terms to shorter JSON keys
07:29:58 [dape]
... use power of semantic web AND being attractive for developers also
07:30:39 [Victor]
the problem here, from the point of view of non-semantic experts, was the lack of a visual representation of the TD model that would be independent of RDF
07:31:52 [Victor]
so, the simplest solution to the issue here is to produce another diagram for WoT developers where only JSON keys are mentioned
07:32:32 [dape]
Maria: unfortunately I could not join the call on Wednesday
07:32:43 [dape]
... so in fact no general agreement
07:33:00 [dape]
Victor: Maria and Fernando described what has be done so far
07:33:13 [dape]
... simplest solution is to have 2 diagrams
07:33:24 [dape]
... keep diagram that is consistent with ontology
07:33:37 [dape]
... another diagram used for Web developers
07:33:55 [DarkoAnicic]
07:33:56 [dape]
... seconds diagram only uses JSON keys
07:34:08 [dape]
ack Victor
07:35:20 [dape]
Fernando: agree that we might want to describe semantic and syntactic world and its association
07:35:35 [dape]
... nevertheless we still think coupling is not good
07:36:53 [dape]
Sebastian shows slide about OCF and oneM2M model
07:37:06 [MariaPoveda]
I thing coupling the naming between ontology and json will limit the reuse of other ontologies, you can not say the others to change their URIs, therefore this coupling strategy seems to have a short proyection IMO
07:37:20 [dape]
SK: would like to point to other standardization groups also
07:37:53 [dape]
Darko: There is disagreement...
07:38:02 [dape]
... so far no agreement
07:38:25 [dape]
... we have different proposal
07:38:35 [dape]
... none solved disagreement
07:39:08 [dape]
... current proposal in favor of web community and not semantic community
07:39:32 [dape]
... what do we need to agree/align?
07:40:09 [MariaPoveda]
07:40:11 [dape]
... let's stay with the name of the thing ... change name for interactions
07:40:39 [dape]
ack DarkoAnicic
07:40:55 [dape]
... think this little change could solve disagreement
07:41:48 [dape]
Maria: I am fine with using "name"
07:42:05 [dape]
... I do have problems with removing links
07:43:10 [dape]
07:43:16 [Victor]
07:43:40 [dape]
... also disagree with outputData and inputData
07:46:50 [kaz_]
(some more discussion on
07:47:39 [kaz_]
07:47:43 [kaz_]
ack m
07:47:45 [kaz_]
ack v
07:49:07 [kaz_]
kaz: suggests we collect requirements for our need
07:49:23 [kaz_]
... and if can be reused, that's fine
07:49:30 [kaz_]
... but if something is missing, we need to add that
07:50:34 [kaz_]
seba: we should follow all the expectations
07:51:03 [kaz_]
... I'd understand what the problem is
07:51:24 [kaz_]
fernando: would describe why mapping would not work
07:51:53 [kaz_]
seba: ok
07:52:16 [MariaPoveda]
As I said there is no technical reason, unless you want to translate the ontology URIs. it is about good practices I already explained in the emails
07:52:58 [Fernando]
In fact, what I proposed Sebastian is to provide a technical reason on why putting mappings in @context will not work in general
07:53:09 [yingying]
yingying has joined #wot-td