Re: BP editors draft ready for review

I’ll try to dial in. Enjoy lunch.

> On Dec 16, 2016, at 8:26 AM, Jeremy Tandy <jeremy.tandy@gmail.com> wrote:
> 
> Thanks Josh. We're picking up on BP in 35 mins when lunch is done. Are you planning to join?
> On Fri, 16 Dec 2016 at 13:25, Joshua Lieberman <jlieberman@tumblingwalls.com <mailto:jlieberman@tumblingwalls.com>> wrote:
> I'm leaning towards "seeAlso" for ISO UML. a little leery of more functional relationships between the OWL and UML entities.
> 
> Josh
> 
> 
> On Dec 16, 2016, at 05:19, <Simon.Cox@csiro.au <mailto:Simon.Cox@csiro.au>> <Simon.Cox@csiro.au <mailto:Simon.Cox@csiro.au>> wrote:
> 
>> For the SSN/SOSA-OM alignment I am suggesting that we use the GOM URIs to support alignment to the UML. Then also do a mapping to om-lite. But that is perhaps a special case where the OGC/ISO predecessor is specific and clear.
>> 
>>  
>> 
>> Simon
>> 
>>  
>> 
>> From: Clemens Portele [mailto:portele@interactive-instruments.de <mailto:portele@interactive-instruments.de>] 
>> Sent: Friday, 16 December, 2016 18:38
>> To: Cox, Simon (L&W, Clayton) <Simon.Cox@csiro.au <mailto:Simon.Cox@csiro.au>>
>> Cc: Linda van den Brink <L.vandenBrink@geonovum.nl <mailto:L.vandenBrink@geonovum.nl>>; public-sdw-wg@w3.org <mailto:public-sdw-wg@w3.org>
>> Subject: Re: BP editors draft ready for review
>> 
>>  
>> 
>> I know, but do we really want to recommend the GOM approach? Is there evidence that the use of these RDF representations are a "best" or even "common" practice?
>> 
>>  
>> 
>> Clemens
>> 
>>  
>> 
>> On 16 Dec 2016, at 00:35, Simon.Cox@csiro.au <mailto:Simon.Cox@csiro.au> wrote:
>> 
>>  
>> 
>> Ø  what would be the advice regarding mapping the types in the ISO schemas to other representations, in particular RDF and JSON?
>> 
>>  
>> 
>> FWIW, an official-ish URI for each element (class, attribute, association role) in the ISO models is provided through the auto-generated OWL representations prepared by the ISO/TC 211 Group of Ontology Management. The RDF representations are available here:
>> 
>>      https://github.com/ISO-TC211/GOM/tree/master/isotc211_GOM_harmonizedOntology <https://github.com/ISO-TC211/GOM/tree/master/isotc211_GOM_harmonizedOntology>
>> For example, the URI for GM_Point is
>> 
>>    http://def.isotc211.org/iso19107/2003/GeometricPrimitive#GM_Point <http://def.isotc211.org/iso19107/2003/GeometricPrimitive#GM_Point>
>>  
>> 
>> However, this does not (yet?) link to the RDF representation (or to anything else). URI not URL.
>> 
>>  
>> 
>> Simon
>> 
>>  
>> 
>> From: Clemens Portele [mailto:portele@interactive-instruments.de <mailto:portele@interactive-instruments.de>] 
>> Sent: Thursday, 15 December, 2016 22:39
>> To: Linda van den Brink <L.vandenBrink@geonovum.nl <mailto:L.vandenBrink@geonovum.nl>>; public-sdw-wg@w3.org <mailto:public-sdw-wg@w3.org>
>> Subject: Re: BP editors draft ready for review
>> 
>>  
>> 
>> Dear all, 
>> 
>>  
>> 
>> on my way to London I have been reading the new draft. Good work! 
>> 
>>  
>> 
>> I have made some edits in a branch (that I will publish later) and list some comments for discussion (in London or in 2017) below. I can add GitHub issues, if this is preferred.
>> 
>>  
>> 
>> a - General: The scope states "this document provides advice on: the choice of ontology and data format to be used when encoding spatial data; …". Since many application schemas exist that use types defined by the ISO 19100 standards and GML data compliant to these application schemas are made accessible in SDIs, what would be the advice regarding mapping the types in the ISO schemas to other representations, in particular RDF and JSON? For RDF probably the work on the ontologies (sensors, time, coverages) will help, but should we provide more explicit guidance somewhere to be consistent with the scope? 13.5 is not complete at the moment.
>> 
>>  
>> 
>> b - BP1: Chapter 9 states as the first reason why SDIs are not enough that the SDI catalog services are insufficient for the general web (which I agree with). But then the first (!) BP is to provide metadata in the dataset metadata in these catalog services (or similar services for DCAT). Seems a bit inconsistent to me.
>> 
>>  
>> 
>> c - BP1: The approach to implementation is heavy on DCAT, but does not mention, for example, schema.org <http://schema.org/> or other ways to represent such metadata. Isn’t DCAT mainly the ontology used by a certain community just like ISO 19115 is the ontology typically used by the geo-community?
>> 
>>  
>> 
>> d - BP2: It would help, if there would be more specific advice how to represent the UoM of a quantity value. Will the SOSA work provide this for RDF representations?
>> 
>>  
>> 
>> e - BP3: As axis order has been and is a common issue, maybe mention it somewhere and include a reference to the OGC Axis Order policy?
>> 
>>  
>> 
>> f - BP4: I could add ldproxy.net <http://ldproxy.net/> as another example and address issue 447 based on our experiences.
>> 
>>  
>> 
>> g - BP4: Currently this is mainly about the spatial things, but the dataset metadata should also be indexed.
>> 
>>  
>> 
>> h - BP4: Sitemaps only work for small datasets.
>> 
>>  
>> 
>> i - BP7: The paragraphs in "Possible Approach to Implementation" are confusing. First the reuse of authoritative identifiers is promoted in the first paragraph and then it is explained why this is in general not a good idea. Why not mainly promote the "mint your own URI and reference the other source" approach and maybe mention the reuse for special cases where it makes sense (whatever these are)?
>> 
>>  
>> 
>> j - 13.5: Why only list RDF vocabs in the note not also other vocabs like 19107, WKT, GML, GeoJSON, etc?
>> 
>>  
>> 
>> k - BP8: Referring to "WGS 84" as a CRS is a bit dangerous. Many CRSs in the EPSG register either have "WGS 84" as name or as part of the name. Maybe clarify up-front that in the document, when talking about "WGS 84" as a CRS, it actually means http://www.opengis.net/def/crs/EPSG/0/4326 <http://www.opengis.net/def/crs/EPSG/0/4326> (or which other variant, lat/long/height, long/lat, x/y/z, etc).
>> 
>>  
>> 
>> l - BP10: The "Possible Approach to Implementation" is very RDF oriented (see comment j above).
>> 
>>  
>> 
>> m - BP13: "protocol independent"? HTTP is a protocol, too!
>> 
>>  
>> 
>> Best regards,
>> 
>> Clemens
>> 
>>  
>> 
>>  
>> 
>> On 13 Dec 2016, at 21:09, Linda van den Brink <L.vandenBrink@geonovum.nl <mailto:L.vandenBrink@geonovum.nl>> wrote:
>> 
>>  
>> 
>> Hi all, 
>> 
>>  
>> 
>> The BP editors draft[1] is ready to be reviewed. As you know we've planned a vote to release a new WD, this Friday. 
>> 
>>  
>> 
>> You should re-read BP4, 6, 7 and 11. These have been significantly updated. 
>> 
>>  
>> 
>> Minor changes include the addition to section 9 of a list of most important BPs when starting from an existing SDI, change of a few BP titles to include the word spatial
>> 
>>  
>> 
>> For all changes since the last WD see [2] (changes after October are new).
>> 
>>  
>> 
>> [1]: http://w3c.github.io/sdw/bp/ <http://w3c.github.io/sdw/bp/>
>>  
>> 
>> [2]: https://github.com/w3c/sdw/commits/gh-pages/bp <https://github.com/w3c/sdw/commits/gh-pages/bp>
>>  
>> 
>> Linda
>> 
>>  
>>