Meeting minutes
Agenda
<kaz> agenda for today
Minutes Review
<kaz> Feb-9
Lagaly: we should not use P1 tag or requirement so that people are not confused
… in future issues/PRs. The list is fixed
definition of OOTBI
<kaz> Issue 155 - Agree on a common definition of "out-of-the-box-interoperability"
Lagaly: Ben has given a definition
… I have also given these four layers which is from a paper that Dave Raggett is an author of
… we have action semantics for example for the case of semantic interop
… we also have a proposal from Cristiano
… some items from cristiano's proposal can map to the other one
Lagaly: ben proposes that a profile should not try to solve all of the above interoperability layers
bf: these layers of definition is good for a general definition, as a reference but we have to discuss whether the profile has to fullfil others
Lagaly: I think it like going to a store and buying 2 devices from 2 manufacturers and they can be hooked up together easily
… for example buying a TV and connecting it to an antenna from another manufacturer and they work
bf: can we a bit more concrete on the layers, michael lagally can you explain what a profile should constrain in each of the layers
Lagaly: semantic interop is agreed on
… organisational means that TDs and profiles should be self descriptive
bf: how about to the 2 first
Lagaly: these are also agreed on
Sebastian: it seems that you are looking at the profile like buying an apple homekit compatible devices and they work together. Instead, this should be not specific to the manufacturer
… I think this is not possible or not possible with WoT
Lagaly: but that is why we have standards, no?
Sebastian: but in this case, W3C is not the right SDO
Kaz: as we discussed during the editor's call yesterday, WoT is a technology for people to handle physical entities in an interoperable manner with each other by an abstraction based on the Thing Description.
… second point, web technology includes Web browser. These days Web browser technology is very popular like TVs, PCs, Smartphones etc. It is popular because it is usable regardless of the platform, hardware, OS
Kaz: Those two points are the more important, and we should concentrate on our own definition of "interoperability" based on our own use cases. So I don't think we have to start with the EU project definition as the basis.
Lagaly: (as a response to SK) This is a strong statement. I disagree that W3C is not the right place
Kaz: I believe Sebastian's point was about concentrating on the abstraction discussion rather than thinking about the details on the physical layer
bf: I am somewhere in between. I would like to see two organizations' devices interoperating. I agree that W3C is not an SDO with a centralized certification. Thus, consumers should be able to handle via fallbacks in cases where the fallback is implementable
Sebastian: I would like to explain my point in a proper manner. In the example of a medical and automotive device, even if they follow a profile, I am sure that they will never be able to talk to each other because they have a different semantic information in the properties
… so the wish is not possible in my opinion
Lagaly: I think that I have another understanding of semantic interop
Lagaly: we have 15 minutes left, let's write the definition
ege: I think that the semantic interop means two things: one is about how we use affordances (sync actions, writable properties etc.), other is about ontologies. I think that sebastian says that the ontology level interoperability is very difficult and I agree. A medical pump can use a different set of units and meaning of affordance names vs a pump in a sewage treatment plant
Lagaly: (writes comment to issue 155) (scribe will not scribe the comment)
Kaz: we should be specific about organizations. One is vendor specific and the other is local SDO.
Lagaly: aob?
<kaz> updated comments on interoperability
<kaz> [adjourned]