Re: Notes on SDWBP - for discussion in 25 June call

Thanks, Bill.

I was about to write a similar email after a scan of the document, so I just add additional thoughts inline.

Am 25.06.2020 um 10:08 schrieb Bill Roberts <bill@swirrl.com<mailto:bill@swirrl.com>>:

Hi all

One of the agenda items for the call today is potential updates to the SDWBP.  Along with a few others I offered to have a look through and make some suggestions for what that might involve.

On reading back through, I was pleased to remember that it's pretty good!  Also in most respects it seems fairly up to date, and I did not find any big issues.

+1 from me. The basic statements are all still valid, it is mainly that we probably want to have some updates, in particular in the examples and notes.


Some specific points:

- check in detail that all links in examples are still valid.  I tried quite a lot of them and haven't found broken ones yet, but I didn't check comprehensively

I have found a few that did not work. For example, the SFPark link in the very first example 1 gives a database error. Some OS links are not longer working. Etc.

- review implementation reports on SDWBP and see if those point to good examples that could be added to some of the best practices

+1, and we could add a link to the implementation report landing page to the front page of the BP document. Maybe we should also try to get more reports and update the existing ones (I have that planned for "my" report, based in and pending update of the APIs).

- add information on the new OGC APIs - section 12.3 and BP23 in particular seems an appropriate place but perhaps also in other parts of the doc

Another place to mention the OGC API development is at the end of section 11.

- in BP1, we could update the link to DCAT to point to the revised DCAT2 instead. Also review DCAT2 to see if any of the new features are particularly relevant for spatial data
- in BP2, we could perhaps add more detail on how to use schema.org<http://schema.org/> metadata to support search engines as schema.org<http://schema.org/> has moved on quite a bit in the last 3 years
- possibly add more explanation or examples of how non-spatial data (whether a web page or linked data) can be connected to/from spatial aspects of a spatial thing.  This is covered in 12.1.3 but the OGC APIs make this kind of thing easier (in comparison to say WFS or typical features of GIS software) and could make for a good illustration

Other ideas for discussion:

- web linking, BP10: the document currently has a bias towards RDF properties; web linking (RFC 5988/8288) and link relation types play a much smaller role, eg, in BP10; I think this should be changed, potentially with a discussion how these relate, which would also be helpful for current discussions related to OGC API (uses web linking) and JSON-LD
- should we add more about 2d/3d "vector tiles" (MVT, 3D Tiles, i3s)? Vector tiles are mentioned here and there, but maybe it deserves a separate paragraph and potentially also some recognition in annex A?
- add something about dynamic CRS / coordinate epoch in section 9 / BP8?
- BP7: add CRS84/CRS84h as alternatives to 4326/4979, add Content-Crs http header as an option how to express the CRS of a response where the encoding has no way to explicitly state the CRS?

Best regards,
Clemens



Talk to you later

Bill