Update rules for logical/control variables

[mkavulich]

Description

Update the rules for logical and control variables to distinguish what prefixes are allowed and when they should be used, based on the proposal in Issue #101. Also update a few standard names based on these rules.

This PR would supersede #149

Issues

Resolves #101

[mkavulich]

@nusbaume When you get a chance, can you look at this and see if it's a good substitute for your #149?

[climbfuji]

I am ok with this. There is always going to be some ambiguity what to use, but hopefully the rules PLUS existing names in the dictionary will help.

[nusbaume]

Thanks for the PR @mkavulich! I'm happy with this replacing my previous PR.

With regards to the standard names, was there a reason why flag_for_ was chosen over is_? I personally find is_ to be clearer, especially when do_ is also allowed (as one is for a given state or condition, while the other is to indicate a specific action, and both make it clear as to what a .true. value means). However I could certainly be convinced otherwise if people have reasons they prefer flag_for_. Thanks!

[mkavulich]

@nusbaume You bring up a good point; is_ is more analogous to do_ in the context of the definitions I proposed in this PR, and also feels naturally less ambiguous. It would also have the advantage of making names shorter. I have updated my PR accordingly.

@climbfuji Since you already provided a review, are you okay with this change?

[climbfuji]

@nusbaume You bring up a good point; is_ is more analogous to do_ in the context of the definitions I proposed in this PR, and also feels naturally less ambiguous. It would also have the advantage of making names shorter. I have updated my PR accordingly.

@climbfuji Since you already provided a review, are you okay with this change?

I don't have a strong opinion on this one, happy with @nusbaume's suggestion. As long as we - at some point - start using those names in the host models!

[nusbaume]

Thanks all! I'm happy with this PR now.