W3C home > Mailing lists > Public > public-sdw-wg@w3.org > March 2017

[public-sdw-wg] <none>

From: Rob Atkinson <rob@metalinkage.com.au>
Date: Mon, 13 Mar 2017 01:51:26 +0000
Message-ID: <CACfF9Lxbmcq1o-KVHSXb0XgyPWZy1Q8asGf8=KeZ+Ryf09CpAw@mail.gmail.com>
To: SDW WG Public List <public-sdw-wg@w3.org>
voting on "module integration"

The comments on the votes suggests there is an issue with the baggage that
goes with various namespace "shorthand". Specifically "SSN" seems to be
used by some people as a shorthand for SOSA axiomitisation and others for
"extended semantics of specialised classes related to sensing activities".
These two usages are mutually incompatible in the context of a discussion
attempting to tease out the separate concerns (even if we end up bundling
the two types of assertions in a single module later)/

I realise that an attempt was made to avoid this by using the "unify"
placeholder - and in retrospect i think i was partly to blame for moving
away from that - looking for a simple concrete option based around existing
namespaces (a minimal tweak).

IMHO all votes should be redone once people get a handle on this issue. I
cannot reconcile current rationales with a consistent sense of what SSN is
supposed to be.

Once the issues are cleanly separated, we can see that a number of options
collapse into each other under certain circumstances:

Option 1 and Option 8 will look the same if SSN introduces no new terms (I
think this might be Kerry's sense of SSN being the axioms for SOSA, and
SOSA having all the terms -a superset -  needed for SSN model)

Option 5 and Option 8 will look the same if we need no extended OWL axioms
for SOSA (which AFAICT is the status quo - SSN is restricted semantics for
sensors, there does not appear to be any SOSA-only OWL axioms anywhere -
though I suggest that Option 8 would allows us to put a placeholder for
these and address them in a followup activity without needing to version
SOSA.

I also think we should look at the use case of someone wanting to further
model a particular class of sensor or  actuator - does the integration
pattern provide a clear example of how to do this?

Rob
Received on Monday, 13 March 2017 01:52:13 UTC

This archive was generated by hypermail 2.3.1 : Monday, 13 March 2017 01:52:14 UTC