W3C home > Mailing lists > Public > public-sdw-wg@w3.org > April 2016

Re: Best Practice sub-team call: 14:00 UTC, 20-April-2016

From: Ed Parsons <eparsons@google.com>
Date: Mon, 25 Apr 2016 11:37:12 +0000
Message-ID: <CAHrFjcnNsMHDyUXsY-tiN8d5ENaBRN0Q9ec-PZOifZSB5vpz9Q@mail.gmail.com>
To: Jeremy Tandy <jeremy.tandy@gmail.com>, Joshua Lieberman <jlieberman@tumblingwalls.com>, Frans Knibbe <frans.knibbe@geodan.nl>
Cc: SDW WG Public List <public-sdw-wg@w3.org>
Frans et al,

Would you be up for discussing this topic on the plenary call this week - I
think it is of wider interest and the GeoSPARQL approach seems to be a good
starting point ?

Ed


On Wed, 20 Apr 2016 at 05:56 Jeremy Tandy <jeremy.tandy@gmail.com> wrote:

> Thanks Josh- good input to the discussion. I note your suggestion of
> "adopting the GeoSPARQL ontology as a best first start [...]"
>
> On Wed, 20 Apr 2016 at 13:53 Joshua Lieberman <
> jlieberman@tumblingwalls.com> wrote:
>
>> Regrets over having a conflict for today’s call. It will be useful to
>> define what is meant or covered by “spatial ontology”. A notion of things
>> that might mean is covered here:
>> http://www.w3.org/2005/Incubator/geo/XGR-geo-ont-20071023/
>>
>> ISO 19107 covers geometries and operators, but the feature model as
>> presented in 19109 is perhaps more relevant to the semantic question
>> whether features are relevant on the Web. My sense is that feature
>> discernment is relevant everywhere but there is resistance to the
>> “complexity” of being explicit about it in Web content.
>>
>> The biggest hurdle may still be the divide between coordinate sequences
>> as literals and coordinates as individual objects. Amazingly, that colors
>> almost every aspect of debates over spatial ontologies, such as whether
>> single latitude and longitude properties are enough  for anything
>> worthwhile on the web.
>>
>> Lastly, “webbiness” is good, but so is geometric and geodetic validity
>> and consistency. The BP group should consider adopting the GeoSPARQL
>> ontology as a best first start and charging OGC with developing a normative
>> update to that. Everyone likes to make their own ontology, but the
>> compatibility and computability issues then make for problems in
>> large-scale spatial data representations.
>>
>> —Josh
>>
>>
>> Joshua Lieberman, Ph.D.
>> Principal
>> Tumbling Walls
>> jlieberman*tumblingwalls.com
>> +1 617 431 6431
>>
>> On Apr 20, 2016, at 5:53 AM, Frans Knibbe <frans.knibbe@geodan.nl> wrote:
>>
>> Hello Jeremy,
>>
>> Thank you for acknowledging the issue. Yes, I can try to introduce the
>> problem in the meeting. If we decide we could try working towards the
>> spatial ontology there are still a lot of ways in which we could do that.
>> To start with
>>
>> the mathematical foundations was just a suggestion. Another step could be
>> to take a good look at ISO 19107 abstract model and see how that fits web
>> requirements and see if it is possible to make that model available as web
>> semantics, to be used as a foundation for other web models involving
>> spatial data.
>>
>> Regards,
>> Frans
>>
>>
>>
>> 2016-04-19 23:28 GMT+02:00 Jeremy Tandy <jeremy.tandy@gmail.com>:
>>
>>> Frans - thanks for the suggestion. Simon, Andrea - thanks for
>>> discussion. Certainly a thorny issue to resolve, but one that we need to
>>> address openly and transparently. Even if a single solution cannot be
>>> found, we _will_ need to document the reasons why a single solution is not
>>> possible - which will help us establish [best] practices for when and where
>>> each of the multiple solutions should be used. Of course, life would be
>>> easier with one solution; so we should begin by striving for that.
>>>
>>> If we're going to start from mathematical foundations, this sounds like
>>> a rigorous piece of work. Would it be best published as an independent
>>> Note? (the WG has the remit to do so if we see fit).
>>>
>>> I've updated the agenda [1] to include the proposed topic (see below).
>>> Frans - can you take the lead on this subject please? Given that this is a
>>> massive topic, we won't finish the discussion in one meeting! So I will
>>> time-box the discussion to 30-mins to allow for time to discuss other items.
>>>
>>> *Part 1 (30-mins): establishing an "agreed spatial ontology"*
>>>
>>>    - *[Frans] problem statement*
>>>    - *solution criteria - what do we need the "agreed spatial ontology"
>>>    to do?*
>>>    - *prior art - what can we learn from (and where are the overlaps
>>>    and gaps in existing work)*
>>>    - *define plan of work - how do we move forward; who will lead?*
>>>
>>>
>>> [1]:
>>> https://www.w3.org/2015/spatial/wiki/Meetings:BP-Telecon20160420#Main_agenda
>>>
>>> On Tue, 19 Apr 2016 at 11:57 Frans Knibbe <frans.knibbe@geodan.nl>
>>> wrote:
>>>
>>>> Yes, it would be tough issue. But I am afraid that not addressing the
>>>> issue properly will incur much greater costs. Perhaps not on our group, but
>>>> certainly on society and the web as a whole. But also within our group
>>>> there can be immediate benefits of having a basic spatial ontology. I have
>>>> the feeling that many of the problems we are trying to solve are are a
>>>> result of the absence of a solid theoretical foundation of the things we
>>>> try to work with. We are mostly scratching the surface instead of attacking
>>>> the core issues. I think it is likely that if the few core issues are
>>>> resolved satisfactorily, many other problems will cease to be problems.
>>>>
>>>> I also think that the undertaking of defining a spatial ontology can be
>>>> broken up in consecutive steps. We do not need to finish all those steps in
>>>> order to be successful. But we can leave a solid foundation for others
>>>> to continue with.
>>>>
>>>> To keep things simple, a first goal could be just to define vector
>>>> geometry as a data type. Having just that would solve a lot of
>>>> interoperability problems and would clear the way for universal ways of
>>>> storing and exchanging spatial data. Humanity has been able to do the same
>>>> for numbers and text, and having agreed upon models for those data types
>>>> makes IT a lot simpler than it would have been if there was no such
>>>> agreement.
>>>>
>>>> So a first step could be defining vector geometry as a mathematical
>>>> construct, an ordered set of coordinates in some reference system. Much
>>>> could be built on top of such a foundation, and even if in the end there
>>>> would still be need for different serializations of geometry, it would help
>>>> if those serializations share a common base model - going to a more basic
>>>> level would achieve interoperability. And if the mathematical foundations
>>>> are solid, a lot of derived ways of working with spatial data
>>>> (transformations, spatial relationships, topology,...) will be much easier
>>>> to set up. And as said before, we would also have a common ground for
>>>> geographic geometry and non-geographic geometry, so that we can use methods
>>>> from both worlds interchangeably.
>>>>
>>>> Perhaps what has been going on in the area of improving temporal data
>>>> on the web can be an example for spatial data. I noticed that when a
>>>> mathematical approach to time is adopted, the difference between time
>>>> instants and time intervals disappears. On the surface there still is
>>>> a difference, but at the core the situation is simpler. At the core, the
>>>> concept of time is more unified than all the different expressions that can
>>>> be encountered in the wild. I think that the same principle - going to the
>>>> core of the matter to make things simpler and more universal - can also be
>>>> applied to spatial data.
>>>>
>>>> Regards,
>>>> Frans
>>>>
>>>>
>>>>
>>>> 2016-04-19 1:36 GMT+02:00 <Simon.Cox@csiro.au>:
>>>>
>>>>> +1 for considering this openly.
>>>>>
>>>>> It is probably the issue that the wider community would most expect to
>>>>> see dealt with in the WG. However, it is definitely a tough issue, and I’m
>>>>> sceptical that it is possible or even desirable to imagine that a single
>>>>> solution is necessary.
>>>>>
>>>>>
>>>>>
>>>>> *From:* Frans Knibbe [mailto:frans.knibbe@geodan.nl]
>>>>> *Sent:* Monday, 18 April 2016 11:42 PM
>>>>> *To:* Jeremy Tandy <jeremy.tandy@gmail.com>
>>>>> *Cc:* SDW WG Public List <public-sdw-wg@w3.org>
>>>>> *Subject:* Re: Best Practice sub-team call: 14:00 UTC, 20-April-2016
>>>>>
>>>>>
>>>>>
>>>>> Hello Jeremy,
>>>>>
>>>>>
>>>>>
>>>>> Could this meeting be an opportunity to discuss the 'agreed spatial
>>>>> ontology' mentioned in the charter (also see this e-mail thread
>>>>> <https://lists.w3.org/Archives/Public/public-sdw-wg/2016Mar/0057.html>)?
>>>>> I feel it could be the most crucial contribution to the data web our group
>>>>> could make, so it would be good to have more clarity on whether and how we
>>>>> wish to pursue this.
>>>>>
>>>>>
>>>>>
>>>>> Regards,
>>>>>
>>>>> Frans
>>>>>
>>>>>
>>>>>
>>>>> 2016-04-18 12:09 GMT+02:00 Jeremy Tandy <jeremy.tandy@gmail.com>:
>>>>>
>>>>> All. For those participating in the Best Practices sub-team, the next
>>>>> meeting is scheduled for 14:00 UTC this Wednesday (20-April).
>>>>>
>>>>>
>>>>>
>>>>> Preliminary agenda is here:
>>>>> https://www.w3.org/2015/spatial/wiki/Meetings:BP-Telecon20160420
>>>>>
>>>>>
>>>>>
>>>>> Please advise if you want to add anything. If you can't make the
>>>>> meeting, please record your 'regrets', else we'll see you there.
>>>>>
>>>>>
>>>>>
>>>>> Regards - Jeremy, Linda and Payam.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>
>> --

*Ed Parsons *FRGS
Geospatial Technologist, Google

Google Voice +44 (0)20 7881 4501
www.edparsons.com @edparsons
Received on Monday, 25 April 2016 11:37:53 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:31:20 UTC