- From: Bruce Bannerman <bruce.bannerman@bom.gov.au>
- Date: Wed, 19 Apr 2017 07:22:11 +0000
- To: Jeremy Tandy <jeremy.tandy@gmail.com>
- CC: "public-sdw-wg@w3.org" <public-sdw-wg@w3.org>
- Message-ID: <1492586531005.15389@bom.gov.au>
Hi Jeremy,
I've just had a quick scan through the rest of the best practice.
This document needs time for a decent sanity check.
It is quite technical and assumes quite a bit of prior knowledge.
I suspect that we may be missing the needs of the targeted audience (that I'm assuming to be a Web Developer).
As an example under 11.4 Parse that, there is the text:
"Imagery formats JPEG [JPEG2000<http://w3c.github.io/sdw/bp/#bib-JPEG2000>] and PNG [PNG<http://w3c.github.io/sdw/bp/#bib-PNG>] can also be coerced to carry data; providing 3 or 4 channels of 8-bit data values."
An end user would need to understand that imagery is typical presented as a set of bands of data, where each band is a segment of the Electromagnetic Spectrum.
Using the example of a set a 7 bands within a Landsat 7 scene, only three bands are typically presented in a JPEG/PNG file as RGB values. These RGB values in the file may actually depict a different band combination from the image, e.g. data from Blue, NIR and MIR bands.
The paragraph goes on to discuss the need to avoid compressed data, but does not discuss lossy and non-lossy compression algorithms.
Lossy compression should be avoided (as stated), but lossless may well be OK.
Though this will also depend on intended use. If the data is just to be used as a 'pretty picture' background then the lossy format is probably OK.
The JPEG2000 image format that is used as an example is typically compressed with a wavelet compression. This is usually a lossy compression, unless the data creator has explicitly created a lossless image.
This is only one example that I was able to quickly find for this email.
=====
What is driving the current rush to finish this work? Is the deadline just arbitrary that someone has pulled out the air?
=====
There is a lot of good work in this document. I'd like too see it undergo a rigorous edit and sanity check.
I will have to leave this version of the document now, as I really do have other priorities that I need to concentrate on.
Kind regards,
Bruce
________________________________
From: Jeremy Tandy <jeremy.tandy@gmail.com>
Sent: Thursday, 13 April 2017 3:54 PM
To: Bruce Bannerman; Tandy, Jeremy
Cc: public-sdw-wg@w3.org
Subject: Re: BP document FROZEN - vote next Monday [SEC=UNCLASSIFIED]
Hi Bruce. This is all good input which I will weave into the draft ... many thanks for your efforts!
Our dates seem to have got crossed though :-)
The "Monday" in the email title was back in March ahead of the TC meeting in Delft! We voted to release that draft anyway.
Currently, we expect to have a _FINAL_ draft (pending editorial fixes etc) ready for vote on 26 April ... however I've just picked up an extra week of week that was unexpected so we may slide that a little.
So-
1/ thank you
2/ there's still time for more :-)
3/ incorporating the outcomes from the CRS thread is on my to-do list
Happy Easter!
Jeremy
On Thu, 13 Apr 2017 at 06:24, Bruce Bannerman <bruce.bannerman@bom.gov.au<mailto:bruce.bannerman@bom.gov.au>> wrote:
Hi Jeremy,
I’m not going to get the time to review this BP within the timeframe required by SDWWG.
I’ll send you what I have to date. You may be able to do something with it.
5. Spatial Things
* I recall that there was some discussion on imprecisely defined locations. Where did we get to with this. Some examples:
* The Mara-Serengeti ecosystem
* The Great Barrier Reef
* Dogger Bank
* West of the Black Stump
* The Sahara
* the coastline
* The Western Australian Greenstone belt
* The Mary River floodplain
* These could also be considered 'Spatial Things’, we just can’t reliably define their locations with crisp vector boundaries.
6. Coverages
* I note the statement about coverages not being a 'spatial thing’.
* This is quite constraining and implies that only vector data can be a 'spatial thing’.
* Following on from my comments in response to 2 above, many of these features are best represented with imprecise ‘fuzzy’ boundaries. A coverage is a good way of handling this.
On 5. Spatial Things and 6. Coverages
* I don’t like the artificial delineation between ‘spatial things’ and Coverages.
* Both are conceptually ‘Spatial Things’
* We are just talking about vector and raster data in different terms.
* A ‘Spatial Thing’ could be represented using both raster and vector representations, depending on the context of what the representation is to be used for.
* This is probably where a discussion on how ‘spatial things’ are conceptualised as ‘objects’/features/coverages etc could be useful. It could bring in discussions on scale, precision, accuracy, intended use etc.
* I do like the the linkage between a real world concept and multiple representations of that concept. It will allow us to overcome a long term issue inherent in spatial data modelling.
7. Spatial Relations
* Topological Relations
* Topology is more about the connectedness of features e.g.:
* Network trace, upstream, downstream
* Left of, right of
* Etc
* The examples given are conflating topology with spatial overlay functions, e.g.
* Intersects, Overpaps, Point in poly, Buffer, etc
* Perhaps a brief discussion on 'spatial joins’ may be also appropriate here, where the attributes of one feature can potentially be assumed by another feature due to their spatial co-location, e.g.:
* 'this building’ is located in 'this census district', therefore I infer that the census attributes for 'this census district’ apply to the residents of ’this building'
8. CRS
* See my comments in the related email. They are still relevant as noted by Byron in his email of 4 April.
10. SDI
* This document takes an overly pessimistic and narrow view of SDIs.
* SDIs are more than just a Discovery Metadata catalogue.
* They are defined as "the base collection of technologies, policies and institutional arrangements that facilitate the availability of and access to spatial data”. See the SDI Cookbook [1], Chapter 1, SDI.
* Discovery metadata records *can* be very useful for determining the content of a given spatial dataset and its intended use. They are very useful to use to determine if it is appropriate to use the data set for a purpose other than that for which it was created.
* What would be useful is if a way could be found to make these metadata records within the catalogues indexable, together with a way of linking the Discovery Metadata record to the Spatial Data Service(s) that provide access to this data. This is where there is considerable potential for linked data. And yes these services could be configured to provide GUID’s for spatial things, just don’t throw the baby out with the bathwater.
Sorry, that is all that I’ve been able to get to.
Kind regards,
Bruce
[1] http://gsdiassociation.org/images/publications/cookbooks/SDI_Cookbook_from_Wiki_2012_update.pdf
From: Linda van den Brink <l.vandenbrink@geonovum.nl<mailto:l.vandenbrink@geonovum.nl>>
Date: Thursday, 16 March 2017 at 20:04
To: "SDW WG (public-sdw-wg@w3.org<mailto:public-sdw-wg@w3.org>)" <public-sdw-wg@w3.org<mailto:public-sdw-wg@w3.org>>
Subject: BP document FROZEN - vote next Monday
Resent-From: <public-sdw-wg@w3.org<mailto:public-sdw-wg@w3.org>>
Resent-Date: Thursday, 16 March 2017 at 20:04
Hi all,
The BP document is FROZEN and ready for people to read/review at http://w3c.github.io/sdw/bp/.
A massive amount of work has been done during this sprint. To summarize:
- Section 8 CRS, BP1, BP3, BP6, BP8, BP9, BP10, BP11, BP14, BP17 updated
- BP2, BP12, BP13, BP15, BP16, Section 12.8 (Dealing with large datasets), Section 14 Narrative removed
A full account of the changes is in the sprint plan[1].
VOTE is scheduled for MONDAY 20-3-2017 during the face to face meeting in Delft.
Special thanks to Andrea, Josh, Clemens, Bill, Ed, and Byron for putting in the hours this sprint!
Linda & Jeremy
[1]: https://www.w3.org/2015/spatial/wiki/Detailed_planning_BP_document#February_-_mid_March_2017:
Received on Wednesday, 19 April 2017 07:22:52 UTC