See also: IRC log
<billroberts> https://www.w3.org/2016/07/27-sdwcov-minutes
<billroberts> PROPOSED:approve minutes of last meeting
scribenick roba
<kerry> +1
<billroberts> +1
<dmitrybrizhinev> +1
<scribe> scribenick: roba
<kerry> scribe: Rob Atkinson
+1
<billroberts> https://www.w3.org/2015/spatial/wiki/Patent_Call
RESOLUTION: minutes approved
<billroberts> https://www.w3.org/2015/spatial/wiki/Meetings:Coverage-Telecon20160810
<dmitrybrizhinev> https://github.com/ANU-Linked-Earth-Data/ontology/blob/master/ANU-LED.owl
<billroberts> sam - we can see you are nearly on webex - dmitry will go first on his part of the work
<dmitrybrizhinev> https://github.com/ANU-Linked-Earth-Data/ontology/blob/master/ANU-LED-example.owl
dmitrybrizhinev: most of the work
of defining a coverage compatible with way datacube
works.
... have defined subProperties of geo:location etc for
cells
... explained another DGGS example
... support SPARQL queries over image metadata
<billroberts> roba: has also been looking at an ontology that builds on QB for spatial concepts
<kerry> roba asks about qb dimensions for basic geospatial concepts -- no problem with you defining lat/long in your own namespace and a broader one might be better
<kerry> dmitry: yes
<kerry> roba: a subproperty of WGS 84 CRS may not be a useful coord system
<kerry> dmitry: yes when there is a tem as advised by BP gropu we should subclass that
<kerry> roba; will suggest something tomorrow
<kerry> billroberts: this is and rdf approach top tiliing, with metadata in rdf, is a way towards metadata discovery.
bill: is an RDF approach to tiling
<kerry> ...are you looking for a rdf version of what is in the image?
kerry cant help scribing :-)
dmitrybrizhinev: SSN has capabilities?
kerry: featureofinterest missing
in example - what is it for coverage
... we should ask Peter Baumann
roba: feature of interest is the thing that has a property whose value is the coverage range (i think)
kerry: deployment metadata would
be useful
... use PROV to talk about processing
what is on coverage BP scope/
<billroberts> roba: feels that current practices for describing coverage metadata is not good enough
roba: suggests we focus primarily
on describing the spatio-temporal aspects of coverages
... including the "range envelope"
sam: RDF QB not designed for
serialising dense grids - but useful for description
... materialising subset over the wire may be useful
though
... experience/experiment? is storing data in HDF
... serving in RDF
use case is also describing any access method - e.g. WCS
roba: have you done anything reasoning over descriptions?
sam: not yet
<billroberts> qxk k
<dmitrybrizhinev> roba: is it reasonable to ask the user "figure out on your own how to write the query that will give you the metadata you want?"
<dmitrybrizhinev> What I want is give me the entire graph of your metadata, but without the metadata, as a thing
<dmitrybrizhinev> roba: What I want is give me the entire graph of your metadata, but without the data, as a thing
<dmitrybrizhinev> kerry: "something like select * ?"
<dmitrybrizhinev> roba: You might be able to find a best practice, but there needs to be a shortcut, a best practice
<dmitrybrizhinev> roba: You might be able to find a pattern, but there needs to be a shortcut, a best practice
<dmitrybrizhinev> billroberts: This is why we need a well defined data model
<dmitrybrizhinev> billroberts: ... so the common tasks for the user are simple
<dmitrybrizhinev> roba: it's nice having a data model with both data and metadata, but it makes it harder to say "give me all the metadata without the data"
<dmitrybrizhinev> roba: so we need a best practice to say that the metadata on its own should be easily accessible
<dmitrybrizhinev> roba: have a data model that can support both description and serialisation
<dmitrybrizhinev> roba: have a look at the existing working group requirements, maybe we need to define a new one
<dmitrybrizhinev> sorry, I keep making errors
<dmitrybrizhinev> billroberts: have a look at the existing working group requirements, maybe we need to define a new one
<dmitrybrizhinev> roba: I'll take an action to write up a usecase for rdf metadata description
<dmitrybrizhinev> roba: I want prov, I want SSN, I want that extensible description
<dmitrybrizhinev> kerry: one of my favourite things about rdf is it doesn't distinguish between data and metadata
<dmitrybrizhinev> kerry: you can treat it as a view of the data
<dmitrybrizhinev> roba: yes I would treat it as a view of the data, but the way to do this is not yet standard
<dmitrybrizhinev> billroberts: please write up examples of some of the things you would want to retrieve
maybe - maybe thats an implementation detail ?
<kerry> +1 and me too
This is scribe.perl Revision: 1.144 of Date: 2015/11/17 08:39:34 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Found ScribeNick: roba Found Scribe: Rob Atkinson WARNING: No "Topic:" lines found. Present: kerry billroberts dmitrybrizhinev Regrets: scottsimmons Found Date: 10 Aug 2016 Guessing minutes URL: http://www.w3.org/2016/08/10-sdwcov-minutes.html People with action items: WARNING: No "Topic: ..." lines found! Resulting HTML may have an empty (invalid) <ol>...</ol>. Explanation: "Topic: ..." lines are used to indicate the start of new discussion topics or agenda items, such as: <dbooth> Topic: Review of Amy's report[End of scribe.perl diagnostic output]