spatiotemporal capabilties of web data sources (was: ISSUE 14: temporal reasoning and relations)

Hello all,

I am branching off from the other thread, because this is about another


> Note also that Hobbs and Pan refer to a ‘temporal reasoner’, since a
> generic OWL reasoner could not support all the required comparison
> operations, particularly where OWL-Time steps outside OWL datatypes.
> Similarly, GeoSPARQL implicitly defines a ‘spatial reasoner’, to manage all
> the functions .

This remark triggered a sleeping thought: How can a consumer know that a
SPARQL endpoint is spatially or temporally enabled, i.e. it supports some
set of functions and datatypes that are defined externally to the SPARQL

I think the proper way to describe the capabilities of a SPARQL service is
to use the SPARQL Service Vocabulary
<> (namespace: sd). It
has a class sd:Function, which can be used to make it known that the
endpoint supports a certain function (e.g. geof:contains). At first glance,
this seems a good way of making the spatial or temporal capabilities of a
SPARQL service known. A user agent could request the service description
and check whether the extended SPARQL functions it intends to use are
supported by the endpoint. It allows SPARQL services to partially support a
specification like GeoSPARQL - not all GeoSPARQL functions need to be

For datatypes, the SPARQL function datatype() could be used to find out if
a dataset uses special spatial or temporal data types. So that seems to be
covered too.

But I wonder if this covers all possible requirements a data consumer might
have with regard to finding out spatiotemporal capabilties of data sources
on the web. Could there be a requirement that we have not identified yet


Frans Knibbe
President Kennedylaan 1
1079 MB Amsterdam (NL)

T +31 (0)20 - 5711 347
disclaimer <>

Received on Wednesday, 12 August 2015 11:32:25 UTC