Meeting minutes
minutes
ege: any problems?
<Ege> https://
cris: are the names going to be corrected?
ege: we don't know how to fix it
mk: we capture more stuff than the other groups
ege: ok, even if in tpac I got the opposite impression
pull requests
PR 2154
ege: based on last week discussion we were trying to reduce the number of times where you repeat the same ip address
… you can use the connection to point to different IPs in the proposal
luca: we have to consider the IP address is sort of a solved problem, you can have local dns or even mDNS. On the other hand, if we have to consider to describe a Thing that can be consumed in different ways (different local or remote ips).
… or you even with different kind protocols
mjk: another pattern in that you can use mixed protocols in different affordances
ege: ok yes, I've prepared an example in the PR
mizushima: we should pay attention on the use case of using different IPs, for example ip4 or ip6
cris: here we are presenting our preferred why of modelling TDs using this new mechanism, but it should always possibile doing it in the old "verbose" way, right?
ege: yes
cris: moreover I think this mechanism could help clients to chose the right form by simply checking once the connect defintions
… and prefer one connection type
mizushima: The issue depends on device implementation and complicated. We should investigate more
ege: the whole goal of this exercise is to see the impact before going to specification
… next in the PR we have an example with a TD that support multiple protocols
… I don't see very often double protocol
… but it happens to have a mixed between modbus and modbus+tpc
… do you think all of these are different patterns?
cris: I agree
… I think it is a reasonable design
<luca_barbato> +1
cris: but now choosing a form is difficult and you can't use indexing
luca: yes you can
cris: right,
… or you could chose the connection
luca: you can always iterate over the connections array
… and also you can use the connection term per afforandace level
… you can also do expansion on demand
ege: finally we have the use case of defaults
qè
cris: I think we should review op as array if we allow that
dape: I noticed that we were discussing some "desired extensions" (connection as array or even form at the affordance level). We just need to be sure to not over complicate everything
… it should not be too complex to generate the extended TD
ege: I agree, I think if we implement till example 3 we are fine.
ege: AOB ?
[adjourned]