Dear All,
Jonathan makes a fair point that there is no such thing as certified CF compliance.
On sample files with interesting errors, the CMIP5 archive provides an interesting range of examples. In the process of trying to provide constructive feedback to the IOOS team I shrank (by reducing lat, long and time axis lengths) and partially annonymised (by overwriting institute, contact and model names) 6 CMIP5 files (
https://drive.google.com/open?id=0B6AWgiTrQ0xFdmdpczZwQjNMRUU ) . Documenting them sufficiently to be useful to users would be more work .. and its not clear how best to document them. Putting extended comments in the NetCDF files would be possible, but it might be easier to link to a web page with information (e.g.
http://cfconventions.org/samplefile/notes/ex003.html) explaining the example and options. It would be possible to add a list of errors and warnings in a global attribute(s).
Perhaps it should be considered standard practise to create a sample file implementing each new feature introduced and one or two versions with errors which highlight concerns raised in dicussions. Once correct example has been ceated, editing one or two attributes to create and invalid version would not be hard,
regards,
Martin
Agree with the overall thread. Having well-documented examples in either sense is a great step.
The examples on the NOAA site (sorry, imprecise reference without my computer handy) showing how to structure and document different kinds of marine observations might be a good start for such a collection.
Crowd-sourcing other submissions might be a nice way to build a very useful resource.
John
> On Jul 21, 2017, at 14:49, Chris Barker <chris.barker at noaa.gov<http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata>> wrote:
>
>> On Thu, Jul 20, 2017 at 6:45 AM, Jonathan Gregory <j.m.gregory at reading.ac.uk<http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata>> wrote:
>> perhaps it would be possible for
>> us to maintain a netCDF file containing an example of every possible
>> violation of a requirement or recommendation in the CF conformance document,
>
> Without thinking it through, I suspect that it is literally impossible to put it all in one file -- aren't there examples of "do it this way OR that way", and "fi this, then do that", that you couldn't have data that both does and doesn't conform in one file.
>
> So a set of files would be the way to go -- and would be easier to manage as well.
>
> Also, expecting that we could have a COMPLETE test set is a bit of a fantasy.
>
> However, a most-cases test set of files would be really useful.
>
> I think the "trick" would be to establish a data structure that maps a particular file with a set of what parts of CF are violated, and what recommendations are violated.
>
> The CF-checkers could then check what they find against that data. And it would be easy to simply update a file and/or drop a new one into the set and the checker would update its tests.
>
> A glance at the Python cf-checker code looks like there's a lot of test fies already in there,
>
> https://github.com/cedadev/cf-checker/tree/master/test_files
>
> though it looks to me (at a really quick glance) like the accompanying "check" files are output from cf-checker -- so maybe not a good way to specify what cf-checker SHOULD return.
>
> And if anyone wants to write a cf-checker in another language, it would be nice to have the "errors" specified in a simple to read file format (JSON?)
>
>> and maybe an example of everything which is described as legal e.g. all the
>> actual examples in the conventions document, and some more.
>
> again, hard to put in one file :-) -- but yes, a good set of "perfect" files would be good, too.
>
> -CHB
>
> --
>
> Christopher Barker, Ph.D.
> Oceanographer
>
> Emergency Response Division
> NOAA/NOS/OR&R (206) 526-6959 voice
> 7600 Sand Point Way NE (206) 526-6329 fax
> Seattle, WA 98115 (206) 526-6317 main reception
>
> Chris.Barker at noaa.gov<http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <
http://mailman.cgd.ucar.edu/pipermail/cf-metadata/attachments/20170721/99e31ac6/attachment.html>
Received on Tue Jul 25 2017 - 06:44:58 BST
