Re: geospatial sparql

On 27.11.19 19:29, Linda van den Brink wrote:

Hi Linda,

> I wanted to point out that in the geospatial community, there is a> standardized extension of SPARQL, called GeoSPARQL (a standard by the>
Open Geospatial Consortium (OGC)).  I’m part of a coordinated effort at>
OGC to gather change requests for this standard and to bundle this in a>
document explaining the benefits of representing geospatial data using>
semantics and graph technologies, and subsequently outlining some>
shortcomings of the existing GeoSPARQL implementation specification>
that, if addressed, would unlock its potential to a greater extent.
I'm very happy to hear that. At a recent meetup in Switzerland we
discussed that GeoSPARQL is another spec we should revisit.

> I do not know if geospatial requirements for SPARQL have at all been
> considered within the SPARQL 1.2 community group. If this is of
> interest, I’m happy to provide a bridge between this group and the
> GeoSemantics group at OGC.

I am not sure if this should be part of the same group, others should
weight in here as well. But I did talk with multiple people about
GeoSPARQL and I would be more than happy to contribute to a list of
things that should be addressed.

Off the top of my head:

- Make this a proper HTML based spec. The current PDF is really
absolutely unbearable to read & search/navigate through. I even did some
desperate efforts in trying to create HTML out of the PDF but that
turned out to be too much work.
- Revisit some constructs. IMO some are overly complicated and should be
reviewed and others do not make much sense, at least I did not get it
why it is done like that. It is quite difficult to SPARQL some of those
so there is surely room for improvement IMO.
- Add lots of tests & examples. I think the main problem of the spec is
that it is not clear enough so vendors implement it differently. In fact
it is really hard to find two stores that implemented the current spec
the same way, which makes queries unportable between stores.

I think GeoSPARQL is super useful but the current spec quality is its
biggest enemy at the moment.

regards

Adrian

-- 
Adrian Gschwend
@ netlabs.org

ktk [a t] netlabs.org
-------
Open Source Project
http://www.netlabs.org

Received on Wednesday, 27 November 2019 12:43:45 UTC