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

QB4ST discussion follow-up - BP challenges

From: Rob Atkinson <rob@metalinkage.com.au>
Date: Thu, 03 Nov 2016 22:09:35 +0000
Message-ID: <CACfF9LxA2VxcVqqxM1t2BqBmoMtfO4TYGes5KKgqcWVzkGJhmQ@mail.gmail.com>
To: SDW WG Public List <public-sdw-wg@w3.org>
Following on from the discussion about the QB4ST[1] on the plenary call
yesterday, a few notes and follow-up challenges for the broader group:

0) Have pushed through formatting fixes to

1) The value of RDF Data cube (QB)[2] itself was questioned, given a low
level of uptake - but it was recognised that some way of providing metadata
to describe spatial data sets as coverages is required. This notion is
prevalent in the UCR [3] and BP [4] documents, but no specific technology
is identified. Looking at the BP use of QB is mentioned 6 times and no
alternative suggested. So the first questions are:

Q1: what are the BP examples for this case?
Q2: what alternatives exist for describing the structure and associated
semantics for sets of instances that are in use? (note that DCAT, PROV etc
do not provide this detail)
Q3: is there any evidence of using OWL (which may be expressive enough) to
define such things in a way which allows consuming such metadata without
navigation of the ful complexity of OWL - and evidence of uptake by
developers who are not semantic web experts?

2) The BP document [ specifically mentions the W3C basic geo vocabulary.
QB4ST shows how the spatial aspects of this, and of data using it, may be
described in a way which can also describe other spatial properties (other
vocabularies) of data. At this stage this seems to be a QB "Measure" case.
Q4: is there a BP example of declaring additional metadata about a
pre-existing vocabulary so its spatial characteristics can be discovered in
a machine-readable form?

3) The challenges of describing spatial dimensions of data warehouses using
gridded coverages in general was raised (Josh) - I have yet to fully flesh
out an example here though the plan is to go to the OGC DGGS group to look
for exemplars  as this is a key concern for their domain. Josh has offered
to help me work through this - others welcome

Q5: Is there any evidence of using W3C-geo or other vocabulary to nominate
a "dimension" - i.e. a cell for which observations are recorded in a
gridded coverage?
(If so, I'll extend the example to include this)

4) The proposed status of QB4ST is a note, since I am not (currently) in a
position to fully test it out across the necessary range of
implementations. In the meantime I welcome offers to co-edit the note, and
especially any implementation testing possible.

Q6: Is anyone intending to, or able to, extend their implementation to
include descriptive metadata, and if so how would they handle the concerns
covered by QB4ST (and QB) more generally?

Cheers
Rob Atkinson

P.S. please check in with the github doc for latest versions:  I'll be
making fairly regular updates as I find my way through the mechanics of the
document template. :-).  I've just cleaned up formatting a couple of things
I noticed during the run-through.

[1] http://w3c.github.io/sdw/qb4st/
[2]  https://www.w3.org/2003/01/geo/
[3] http://w3c.github.io/sdw/bp
[4] https://www.w3.org/TR/sdw-ucr/
Received on Thursday, 3 November 2016 22:10:24 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:31:27 UTC