RE: Implementation evidence - Producers

13 classes

[cid:image001.png@01D2E101.52A5F330]

23 properties
[cid:image002.png@01D2E101.52A5F330]

From: Krzysztof Janowicz [mailto:janowicz@ucsb.edu]
Sent: Friday, 9 June, 2017 03:07
To: Dan Brickley <danbri@google.com>; Joshua Lieberman <jlieberman@tumblingwalls.com>; Phil Archer <phila@w3.org>
Cc: Armin Haller <armin.haller@anu.edu.au>; public-sdw-wg@w3.org
Subject: Re: Implementation evidence - Producers

I had a quick chat with Phil Archer on these issues recently, and he suggested that there is value in having two different efforts at least share broadly the same approach to the main entities and relationships, even if the type and property names differ. At Schema.org our current interest in sensors is more from the "internet of things" angle than from a "Spatial data on the Web" agenda, although clearly there are strong overlaps. We're also trying to figure out how to bridge to things like QUDT and Data Cube, so it's unlikely we'd adopt all of SOSA wholesale.

I do not know any details of your roadmap but I am a bit more optimistic. SOSA also focuses on the Internet Of Things and is by far not SDW specific (in the sense that you can make observations about any topic but all of them are, of course, in space-time). Our examples range from smart homes to measures of water quality by government agencies. SOSA is the simplest model that (somewhat) completely captures the observation process we were able to come up with and agree on. I would not expect that schema.org has interest in all the SOSA classes and properties but I strongly believe that there will be a common (and alignable) core. After all, there is no way around talking about sensors, observations, the observed properties, and the things being observed. It would be fantastic if schema.org would reuse some of our work, but most importantly both should be alignable to foster interoperability. The only scenario, I would hope is preventable would indeed be two parallel approaches. This would also undermine the entire point of having a consensus-based standard in the first place. IMHO, seeing a wide adoption of such a joint OGC/W3C standard is not only important for publishing, retrieval, and reuse of observations on the Web but also because this was truly an open, bottom-up effort to which a wide set of parties contributed.

[And again, I am speaking for myself here, not necessarily being able to reflect the views of the entire group].

Jano


On 06/08/2017 08:15 AM, Dan Brickley wrote:
On 8 June 2017 at 16:06, Joshua Lieberman <jlieberman@tumblingwalls.com<mailto:jlieberman@tumblingwalls.com>> wrote:
Dan,

It might help for the group to endorse a specific path or procedure for SOSA - in - schema.org<http://SOSA-in-schema.org>, perhaps adoption as opposed to derivation. It would be unfortunate if users had to specify the “W3C” or the “Schema.org<http://Schema.org>” version of a SOSA concept.

That's what can make these situations somewhat frustrating. In retrospect we would've had less trouble for Schema.org's approach to Dataset description if it had been more radically different to DCAT. Instead we adopted a proposal that tried to stay as close as possible to the graph structure from DCAT - https://www.w3.org/wiki/WebSchemas/Datasets - and it was perceived by some as something of a fork.

I had a quick chat with Phil Archer on these issues recently, and he suggested that there is value in having two different efforts at least share broadly the same approach to the main entities and relationships, even if the type and property names differ. At Schema.org our current interest in sensors is more from the "internet of things" angle than from a "Spatial data on the Web" agenda, although clearly there are strong overlaps. We're also trying to figure out how to bridge to things like QUDT and Data Cube, so it's unlikely we'd adopt all of SOSA wholesale.

Would it be useful to commit to making sure there are RDFS/OWL assertions (equivalence, subtype etc.) in any term definitions that can usefully be mapped? Phil - can you offer any guidance?

Dan



—Josh

On Jun 8, 2017, at 10:57 AM, Dan Brickley <danbri@google.com<mailto:danbri@google.com>> wrote:


Thanks - it would certainly be helpful if the WG could include some resolution in its minutes, expressing support for the idea of other schema efforts (such as schema.org<http://schema.org/>) basing their designs partially or fully on SOSA (and SSN?). From the schema.org<http://schema.org/> side we'll certainly explore this enthusiastically, but it would be nice to be officially encouraged :)

On 7 June 2017 at 03:49, Armin Haller <armin.haller@anu.edu.au<mailto:armin.haller@anu.edu.au>> wrote:
I cannot speak for the whole group either, but similar to the feedback from some of my co-editors, the intention of our modelling choices in SOSA was that the core will eventually be “absorbed” into schema.org<http://schema.org/> for maximum impact. I think with proper attribution as was done with GoodRelations there shouldn’t be any concern from the working group. However, I will ask to put that on the agenda for our last SDW general meeting.

From: Dan Brickley <danbri@google.com<mailto:danbri@google.com>>
Date: Wednesday, 7 June 2017 at 10:19 am
To: Armin Haller <armin.haller@anu.edu.au<mailto:armin.haller@anu.edu.au>>
Cc: "public-sdw-wg@w3.org<mailto:public-sdw-wg@w3.org>" <public-sdw-wg@w3.org<mailto:public-sdw-wg@w3.org>>
Subject: Re: Implementation evidence - Producers


I'd be happy to explore that, but it would be good to have some (even informal) encouragement from the SDW WG that this would be a welcome development. We did something similar with a DCAT-based Dataset design a while ago and it seems that we inadvertently caused some frustration as it was seen as a fork. It would be a pity to try to adopt a design based on SOSA (or SOSA/SSN) and for that to be seen in a similar way. Do you think the WG would welcome such an effort?

Dan

On 7 June 2017 at 00:37, Armin Haller <armin.haller@anu.edu.au<mailto:armin.haller@anu.edu.au>> wrote:
Hi Dan,

SSN http://w3c.github.io/sdw/ssn/ is now in rec-track and we are to collect implementation evidences in the next 4 weeks. In fact, we need to show demonstrated use in at least two producer implementations and two consumer implementations, with producer implementations, according to our interpretation of the Director’s words meaning that there needs to be two ontologies that extend each term of SOSA/SSN.

You indicated earlier in the working group that you would likely take the simple SOSA core and integrate all or parts of it in Schema.org<http://Schema.org>. Would that be something you could kick off in the next 4 weeks with maybe a Change request to schema.org<http://schema.org/>?

Cheers,
Armin









--

Krzysztof Janowicz



Geography Department, University of California, Santa Barbara

4830 Ellison Hall, Santa Barbara, CA 93106-4060



Email: jano@geog.ucsb.edu<mailto:jano@geog.ucsb.edu>

Webpage: http://geog.ucsb.edu/~jano/


Semantic Web Journal: http://www.semantic-web-journal.net