See also: IRC log
<billroberts> do please join the webex if you haven't already - details on the meeting page https://www.w3.org/2015/spatial/wiki/Meetings:Coverage-Telecon20160601
<scribe> scribe: eparsons
<billroberts> https://www.w3.org/2016/05/18-sdwcov-minutes
billroberts Minutes of last meeting
<kerry> +1
<scribe> Topic : Approve last week's minutes
<Duo> +1
<billroberts> +1
RESOLUTION: Approve last week's minutes
<billroberts> https://www.w3.org/2015/spatial/wiki/Patent_Call
<billroberts> https://www.w3.org/2015/spatial/wiki/Meetings:Coverage-Telecon20160601
billroberts first item last call update
billroberts Landsat data in RDF, update from Reading on coverage json, roba data cube talk
<billroberts> https://www.w3.org/2015/spatial/wiki/Coverage_draft_requirements
billroberts kerry & maik thanks for comments on this
kerry not complete - will do more
billroberts more comments on reqs welcome
kerry Req alignment with others should not be there ?
other standards that is
<billroberts> this was: 5.23 Spatial data modeling issues solved in existing models shall be considered for adoption, e.g. O&M, SoilML or the OGC coverage model.
kerry Add to plenary meeting agenda ?
billroberts sounds like good idea - generic afterall
kerry maybe others like that ?
billroberts Multi-language for example ?
kerry - No :-) req for ontology developed here needs language options
billroberts Go through at look for other generics to discuss at plenary
<billroberts> https://www.w3.org/2015/spatial/wiki/Candidate_high_level_technical_approach
billroberts link is attempt to bring together work so far from others
billroberts Need some metadata for extracts + actual data (domain)
billroberts metadata in RDF
billroberts Real data in more compact form - e.g. json array ?
billroberts or image data
billroberts RDF metadata - use json-ld for context
billroberts supply metadata first and then data 2 steps
billroberts Different levels of metadata for collections of objects
billroberts Should in terms of scope select only some use cases
<billroberts> https://www.w3.org/2015/spatial/wiki/Candidate_high_level_technical_approach
roba list is better than req doc - 3 concerns relate to information needed - where is the metdata ?
rob - attached to service, dataset, element etc ?
roba - need to be consistent - currently different in different places
roba - also whats the best encoding ?
roba - 3 separate concerns more accessible to people
roba Different viewpoints result in different metadata
roba Find way to express expectation from different viewpoints
roba Datacube approach leads to RDF working well for metadata
roba - Hard to know scope currently
billroberts Fair point - jumble of req at the moment - open to ideas to bring structure
billroberts Question ? 2 stage approach to retrieve extract - 1) metadata 2) payload
billroberts Want to know what you are reviving first ...
billroberts Makes sense ??
kerry - thinks important not sure yet ...
billroberts Common with geo data - geometry in GML for example
billroberts Need to know if the extract is the one you want and how to deal with it...
kerry Philosophical point of difference between data and metadata ...
kerry Don't wants thinks to be seen as different docs ?
Maik I agree OGC for example as coverage doc - coverage json does not have distinction
billroberts based on Maik approach range has separate id not in same json file
Maik separate id's don't break linking ?
roba Data & metadata as single entity but there might be a metadata view of the entity
roba trick is do id of subsets
roba services are key to this
roba richer encoding makes this work
ByronCinNZ Metadata can be both with data and separate - should support both views
kerry analogy Dublin Core & RDF comes from html meta tag worldview - which failed !
billroberts See data as one big graph separation from a practical perspective to allow delivery of data
Maik What about coverages in different formats, so different metadata embedding ?
billroberts Our object is to find one way or a family of ways to deliver coverages ?
roba Search for a item on ebay, search engine provides ebay object but with links to other things aswell
roba Metadata may be v.large graph needs different views in that case - one view for simple.. another view more complex - nice if these consistent
roba Are there canonical views e.g Dublin Core ? - recommended views e.g dimensions ?
billroberts useful discussion thank you..
billroberts Full spectrum of coverage data complicated , don't need to do all - but key are gridded coverages & point clouds x,y,z,t etc
<kerry> +1
billroberts Question - should we aim just at these ?
+1
<Duo> +1
<ByronCinNZ> +1
<Maik> +1
<roba> +1
billroberts Refining that a bit - regular grid only ?
roba One case alone dangerous from a info modelling pint of view
kerry opposite view - need to focus gridded cases dominate, point clouds maybe but no more..
<roba> feature,t
billroberts Grids can do time series... degenerate case
Maik agree x,y,z,t OK not composite x,y vectors for example
billroberts Makes sense to be extensible, optimised for point clouds later... built on grids
<kerry> +1
roba Time dimension makes all degenerate grids ?
roba Hierarchical best practice is what I'm looking for
billroberts special case that we are trying to solve > million more points
billroberts Time nearly up... will try to work on a updated versions of doc
billroberts improve requirements based on this discussion
<roba> Time dimension makes all degenerate grids => using patterns i time dimension for x,y would address regular grids
billroberts Need to work between calls