Re: ACTION-96 linking to related identifiers

Thanks Jeremy - I think you've listed the most important aspects.  One
potential additional best practice for consideration might be a
recommendation to data publishers to provide some form of
search/reconciliation API, particularly important with non-guessable URL
patterns.

On 2 December 2015 at 13:23, Jeremy Tandy <jeremy.tandy@gmail.com> wrote:

> Hi Bill, Jon ...
>
> Great content along with some very useful examples that we (BP editors)
> can incorporate.
>
> I think that the subject boils down to two best practices ...
>
> From Expressing spatial data
> <http://w3c.github.io/sdw/bp/#bp-expressing-spatial> we have Best
> Practice 13: Assert known relationships
> <http://w3c.github.io/sdw/bp/#semantic-rels> which *will** say something
> along the lines of "if you know some relationships between (spatial) Things
> then publish them - because it's hard to figure out relationships from
> scratch" as your examples illustrate.
>
> And From Linking Data <http://w3c.github.io/sdw/bp/#bp-linking> we have Best
> Practice 20: Provide meaningful links
> <http://w3c.github.io/sdw/bp/#meaningful-links> (include the right
> semantics), Best Practice 21: Link to spatial Things
> <http://w3c.github.io/sdw/bp/#link-to-spatialthings> (link up the Things
> rather than the information objects that describe them e.g. geometry
> objects) and Best Practice 22: Link to resources with well-known or
> authoritative identifiers
> <http://w3c.github.io/sdw/bp/#link-to-auth-identifiers> (reference other
> people's well established resources & identifiers thereof). The middle
> one of these needs some work methinks because it's clearly useful to link a
> Thing to its geometric description ... but we want to create a network of
> related resources using the identifiers for the Things.
>
> * "will" say ... because I've not finished writing things up yet :-)
>
> Thanks Bill. Jeremy
>
> On Sun, 29 Nov 2015 at 16:11 Jon Blower <j.d.blower@reading.ac.uk> wrote:
>
>> Hi Bill, all,
>>
>> Just wanted to say that I found this to be an extremely helpful and
>> informative post, thanks!
>>
>> the BP document might be able to help by categorising some of the most
>> common relationships and perhaps suggest examples of appropriate matching
>> vocabulary terms.
>>
>>
>> Yes, I agree. Some of these issues are very characteristic of spatial
>> data and bang in scope for a BP document I think. We often see abuse of
>> owl:sameAs when a weaker term would be more appropriate. Enumerating the
>> options and use cases would be very helpful.
>>
>> (This has particular local relevance to us here - the University of
>> Reading is actually mostly in the Wokingham district, although most people
>> would still refer to it as part of the Reading urban area. “Colloquial
>> Reading” is different from “administrative Reading”, as it is in probably
>> most cities.)
>>
>> Cheers,
>> Jon
>>
>>
>> On 26 Nov 2015, at 18:29, Bill Roberts <bill@swirrl.com> wrote:
>>
>> Hi BP-editors
>>
>> Here are some initial thoughts on the issues of linking from your own
>> Spatial Thing to other identifiers for the same thing or related things.
>>
>> This action is to expand the text in section 7.2 of the BP draft that
>> currently says:
>>
>> "it's useful to have hyperlinks to things like Geonames, wikipedia, OSM
>> etc (see list on the mailing list, keyword: stamp collecting)"
>>
>> As per http://www.w3.org/DesignIssues/LinkedData.html item 4, it's
>> useful for people to link their data to other related data. In this context
>> we're most frequently talking about either Spatial Things and/or their
>> geometry.
>>
>> There are many useful sets of identifiers for spatial things and which
>> ones are most useful will depend on context.
>>
>> I think there are two main challenges here - discovering relevant URIs
>> that you might want to connect to, deciding what is the nature of the
>> relationship between your original URI and potential link targets, and then
>> finding an existing vocabulary term that accurately reflects that
>> relationship.
>>
>> As an example, let's take Edinburgh. In some recent work with the
>> Scottish Government, we have an identifier for the City of Edinburgh
>> Council Area - i.e. the geographical area that Edinburgh City Council is
>> responsible for:
>>
>> http://statistics.gov.scot/id/statistical-geography/S12000036
>>
>> (note that this URI doesn't resolve yet but it will in the next couple of
>> months once the system goes properly live)
>>
>> Here are some identifiers for Edinburgh and/or information about it that
>> we might want to link to, together with notes about how I found out about
>> them.
>>
>> http://statistics.data.gov.uk/id/statistical-geography/S12000036
>>
>> My identifier is directly based on this one, but the Scottish Government
>> wanted the ability to create something dereferenceable, potentially with
>> additional or different info to the data.gov.uk one.  We're happy these
>> two are owl:sameAs
>>
>> https://en.wikipedia.org/wiki/Edinburgh
>> Found by a google search for Edinburgh site:wikipedia.org).  This is a
>> page about a closely related but perhaps less specific concept of the
>> place. Possible document vs thing distinctions to be made here.  Possible
>> relationships: rdfs:seeAlso, schema:sameAs ? foaf:page?
>>
>> http://dbpedia.org/resource/Edinburgh
>> I know the pattern for changing a wikipedia URI into a dbpedia one, so
>> found it that way.  Relationship: "more or less the same as" but not sure
>> I'd want to go as far as the strict semantics of owl:sameAs
>>
>> http://data.ordnancesurvey.co.uk/id/50kGazetteer/81482 (Edinburgh)
>> Found by OS gazetteer search service for 'Edinburgh' then checking the
>> labels of the results that came up.  OS give it a type of 'NamedPlace' and
>> give it some coordinates.
>>
>> http://data.ordnancesurvey.co.uk/id/50kGazetteer/81483 (Edinburgh
>> airport)
>> Also found by the same OS gazetteer search service for 'Edinburgh'.  This
>> is clearly not the same as my original spatial thing, but I might want to
>> say something like 'within' or 'hasAirport'.
>>
>> http://data.ordnancesurvey.co.uk/id/7000000000030505
>> Found by a search for 'Edinburgh' in the OS 'Boundary Line' service that
>> contains administrative and statistical geography areas in the UK.  The
>> first results of the search were parliamentary constituencies - had to
>> scroll down and look for one that had a stated rdf:type that matched what I
>> was looking for.  It's probably safe to say my identifier is owl:sameAs
>> this one.
>>
>> http://sws.geonames.org/2650225/
>> Found with the Geonames search service:
>> http://api.geonames.org/search?name=Edinburgh&type=rdf&username=demo
>> Once you have found a place in geonames, there are other useful services
>> to find things that are nearby etc. Not sure exactly what this is, though
>> it has a RDF type of http://www.geonames.org/ontology#Feature
>>
>> http://www.openstreetmap.org/relation/1920901  (administrative boundary)
>> machine readable data:
>> http://www.openstreetmap.org/api/0.6/relation/1920901
>> Found via the search box at www.openstreetmap.org.
>> see also
>> http://nominatim.openstreetmap.org/details.php?place_id=127903534
>> and http://www.openstreetmap.org/node/17898859 (node - somewhere around
>> the centre of Edinburgh)
>> I'm not sure of all the options with OSM - I'm sure others in the WG know
>> more -but it has identifiers for nodes, ways and relations, though it seems
>> that these identifiers tend to change quite frequently as the map is edited.
>>
>> The outcome of this example is that it takes a bit of prior knowledge and
>> intelligent manual guesswork to find related URIs.  Some services, eg OS,
>> have useful search facilities, but the results may still need some
>> interpretation. Recommending some standard approach to providing a search
>> facility (or 'reconciliation API') for a collection of spatial data might
>> be a useful best practice.
>>
>> Working out how to accurately describe the relationship is hard in
>> general and the BP document might be able to help by categorising some of
>> the most common relationships and perhaps suggest examples of appropriate
>> matching vocabulary terms.
>>
>>
>>
>>
>>