Meeting minutes
Minutes review
<kaz> Jan-8
Ege: last week we discussed binding templates and uri schemes
… any problems ?
Ege: minutes approved
TD
QueryAction operation
<kaz> Issue 2068 - Specify the result of queryaction operation
Ege: this is a big discussion (at lest this is what we agreed last time)
… to really solve it now (creating a feature for TD 2.0) would take to much time
… how do we want proceed?
… my proposal is to have a dirty solution for TD 1.1
… and then have a real solution later on as part of manageable affordances
Daniel: from my point of view, we need to fix in TD and then see if it works in Scripting
Ege: really fixing in TD 1.1 we would need actually to create an errata
Daniel: I don't like to have an hacky solution
… we don't want to have an errata
Cristiano: I think we can work with it leaving underspecified but without taking assumptions (do workarounds)
Kaz: I agree, if we want to update TD 1.1 specification that would be a normative change and it should be applied to 2.0 (as a new version).
Daniel: What we shouldn't do is creating a method call that needs to be reviewed in the next iteration
… this is more a Scripting API discussion
Kaz: There are two questions here, one about TD and another about Scripting API, so we should clarify two resolutions for them separately.
<EgeKorkan> proposal: The issue 2068 will not be addressed in TD 1.1
RESOLUTION: The issue 2068 will not be addressed in TD 1.1
<EgeKorkan> proposal: In TD 2.0, the issue 2068 will be fixed
RESOLUTION: In TD 2.0, the issue 2068 will be fixed
URI Scheme
<EgeKorkan> node-wot PR 1335 - opcua: nsu in nodeId (fixing #1334)
<EgeKorkan> wot-thing-description Issue 2069 - URL Scheme Decision
Ege: we should use the issue above
… to discuss further
… basically since in OPC UA there could be reserved characters in the nodeId
… the idea is to encode it only when it is needed
… fore example if a node id is http://.... the // won't be escaped
… this is the decision taken for opc-ua
… but there might be implications also for other protocol bindings
Koster: we also discussed to use the path
… with mqtt we are using other parameters
… but I'm not sure if it was impacted
Cristiano: we went back and forth but we decided to use the query parameter which need less escapes
Daniel: what the decoder should do?
… it just need to solve the `#` and `&`, right?
… probably you should encode also `%`
Cristiano: yeah
Daniel: then it means that in the encoding side we are doing something special and in the decoding side not
Cristiano: right that's why it would be better just to note that the nodeId goes in an URI so any special character should be encoded
Daniel: I like this point, ZIP does something similar too. They just define the decoder rules.
Ege: I would just direct people to not encode everything
… because we get better readability
<kaz> Percent-encoding
Kaz: when we say encoding and decoding, does that mean Percent-encoding ?
Ege: URL encoding?
Kaz: yes, they are the same concept
Kaz: We should ask TAG for advice
… when do we want to do that?
Ege: since the discussion is almost settled down I would use issue 2069, summarize the decision, and ask for help
<EgeKorkan> wot-thing-description Issue 2069 - URL Scheme Decision
<kaz> wot-binding-templates Issue 292 - MQTT URI Design
<EgeKorkan> w3c/
Ege: in mqtt there is the same issue
… there we have `#` and `*`
… the decision so far was similar to what is now decided in opc-ua
Ege: probably it was more delicate in mqtt
Cristiano: yes, probably because we were trying to fit the topic in the path element of an uri
Ege: . and .. are problematic in MQTT
… I hope they are not used that much
Cristiano: When `..` and `.` are problematic?
Ege: it seems everywhere
Daniel: we can still use the query parameter
… without having to worry that much about / and URI arithmetic
… and maybe more consistent
Cristiano: having the topic in the path would allow to do some simplification when writing the URL
Ege: ok thank you, tomorrow we will discuss registry inclusiveness
Ege: please be there we want to reach a resolution
[adjourned]