remove "missing value" attribute deprecation in Appendix A

[caron]

Appendix A has for "missing_value" attribute :

"A value used to represent missing or undefined data (deprecated by the NUG)"

should be:

"A value or values used to represent missing or undefined data."

We missed this in TRAC ticket #58.

[Dave.Allured]

"value or values". Since this will be the first time that the possibility of multiple missing_values is made explicit in CF, I recommend that the definitive language from NUG appendix B be included as follows.

CF version 1.5, section 2.5.1, "Missing Data"; add a new paragraph following current paragraph 3, to read as follows:

The missing_value attribute can be a scalar or vector containing values indicating missing data. These values should all be outside the valid range so that generic applications will treat them as missing.

[caron]

Im not crazy about the use of "valid range", and wouldn't recommend using both valid_range and missing_value attributes on the same variable. So from my POV, the NUG language is not great.

So I would prefer just the sentence:

The missing_value attribute can be a scalar or vector containing values indicating missing data.

[Dave.Allured]

Replying to caron:

... The missing_value attribute can be a scalar or vector containing values indicating missing data.

Still for new paragraph in 2.5.1? Agreed.

[biard]

Any reason this hasn't been accepted and resolved? I'd say the last comment by John Caron sounds great.

[jonathan]

I'm happy to support this too. I think it went no further because it didn't have a moderator to push it along. Also, it had not reached the level of support required by the rules, but it does if I support it too!

Therefore should be regarded as accepted, and included in CF 1.7. Thanks to all contributors.

Jonathan

[taylor13]

Just to add an "extra" vote, I'm also in favor of the change. Thanks for following up on this. Karl

[rhattersley]

Applied via ​https://github.com/cf-convention/cf-conventions/pull/66.

NB. The changes in the pull request are only those shown in the "Description" at the head of this ticket - they do not include the new paragraph suggested by Dave Allured.

[markh]

merged onto github master