Conformance: update intro sentence to clarify reason for lack of section 1

[sadielbartholomew]

Topic for discussion

Last time I checked the Conformance Document I noticed (for the first time) and wondered why it began the numbered sections at '2.1 Filename' without anything for a section 1. As @cofinoa has clarified, it is implied by one of the introductory bullets, namely:

The following is a list of requirements and recommendations for a CF conforming netCDF file. They are organized by the section of the CF document that they pertain to.

where the first section of the main document forms the introduction and contains nothing that could be translated into a statement of conformance. Hence no section 1. There are other sections which also lack mapped conformance items and therefore do not appear too, but the first is the most obviously lacking.

We agreed it isn't immediately clear about the reason for the missing sections and that something more explicit could prevent any confusion on such matters.

Therefore I suggest we update the second bullet somewhat, original as above, to clarify. Something like the following seems sensible to me (changes in italics):

The following is a list of requirements and recommendations for a CF conforming netCDF file. They are organized by the section of the CF document that they pertain to, where some sections do not translate to any conformance statements and therefore do not appear below, such as section 1.

(The final clause 'such as section 1' could be left out here but might be useful to illustrate a concrete case).

Alternatively I suppose we could have a section title for all sections with those without conformance items stating 'No conformance statements for this section' or similar, but I much prefer the above solution which seems cleaner (not to mention less work to implement).

[ethanrd]

Alternatively I suppose we could have a section title for all sections with those without conformance items stating 'No conformance statements for this section' or similar, but I much prefer the above solution which seems cleaner (not to mention less work to implement).

I lean in the other direction as I think a Chapter 1 heading with text stating 'No conformance statements ' would be much more visible than adding to the first intro statement (which I would probably skim over, but I probably skim too much).

I don't think going deeper than the chapter level would be needed (and it would be messy and more work, as you say).

It looks like chapters 1 and 9 are the only chapters without conformance statements. Which raises a question for me:

• Are there really no conformance statements in Chapter 9 DSGs? Or was adding them overlooked at the time? I haven't looked too carefully but Section 9.4 "The featureType attribute" appears to have two possible conformance statements: 1) a featureType global attribute is required or recommended and 2) the value of a featureType attribute must be one of the string values listed in Table 9.1.

[larsbarring]

Yes, I agree with @ethanrd regarding chapter 1

[sadielbartholomew]

Fair enough - either way works well I think and given Lars' comment it seems we are 2 vs 1 so far on preferences for which approach to use to manage it, so pending any further opinions let's go with the "heading with text stating 'No conformance statements'" way. I absolutely agree on:

I don't think going deeper than the chapter level would be needed (and it would be messy and more work, as you say).

• chapter level is sufficient IMO also.

Very good question raised, as well - looks like Ros can clarify as per David's comment below.

[cofinoa]

I agree. I am happy to go with the chapter-level stating “No conformance statements”. Chapter level is sufficient in my view as well.

[davidhassell]

I think that Chapter 9 conformances have been written, but haven't yet made it in ... is that right, @RosalynHatcher?

[JonathanGregory]

Thanks for asking the question, @sadielbartholomew. I agree with @ethanrd's suggestion that it's more obvious to have a section 1 which says it has no contents, like "This page is intentionally left blank".

As far as I know, we have never written chapter 9 of the conformance document. Of course, it should have been done years ago. I have had it on a list of things which ought to be done for a long time! In addition to what @ethanrd said, there is a requirement that each data variable has the coordinate variables which it must have according to Table 9.1 for the featureType, and there are various detailed requirements about the ragged array representations (to check they're correctly constructed). Probably there are other things too. It requires a detailed reading. Is there anyone who would be willing to make a first draft, unless @RosalynHatcher has done it already?