[CF-metadata] proposed rules for changes to CF conventions

[John Caron]

Jonathan Gregory wrote:
> Dear John
> 
>> To clarify, my intention was to specify that we need 2 (or more) independent confirmations of the feature with working code, preferably this would be libcf and the CF-checker. (not libcf and the CF-checker plus 2 more)
> 
> OK, sorry, I misunderstood.
> 
>> I am concerned about 1) libcf and the CF-checker being the bottleneck for proposals, especially complex ones that the developers of libcf and the CF-checker might not have time to work on; and 2) libcf and/or the CF-checker would not be an adequate test of the feature.  In both cases it seems useful to allow other software to fulfill the implementation requirement.
> 
> Yes, that is fair enough. I think that any two applications would be
> acceptable. As a proviso, we could say that the cttee is the judge of the
> suitability. I put libcf and cf-checker because they expect to be updated
> but, as you say, it might sometimes turn out to cause a delay.

Yes, the committee will have to decide if the implementations are a sufficient test.