[macula]

I don't know if this has been brought up already in this thread—if so, please disregard and accept my apologies.

It seems to me that the source of all evil is the ability bestowed upon the user to keep a parent "pending" (i.e. not ticked off) while all its children are completed—I'm referring to that insidious option at the very bottom of Preferences>Data. That's how the dilemma that the Omni developers faced emerges: Shall the action following the parent be released anyway (whereby the parent risks "falling through the cracks") or kept unavailable until the parent is ticked off (whereby your project risks stalling with the user possibly in oblivion or confusion)?

It is reasonable and laudable that Omni tries to address this problem. However, the solution they are suggesting in 1.8a is, as many of us in this thread believe, at least as confusing and introduces lots of conceptual and usability problems.

So much confusion, and so much debate pitting two imperfect approaches against each other, just in the name of "flexibility": Allowing the user to keep a parent open even if all its children are completed. In my view, this is too high a price to pay for a damn check box in the preferences!

Here is my somewhat radical suggestion then: Remove the check box altogether and hardcode this behavior whereby a parent will be automatically marked as completed when its children are completed. Then also remove all this "1.8 baggage", most of all the assignment of contexts to parents, which has sparked so much lively debate among both devoted and new users.

The cost will be minimal: If you want to add a further action to a completed parent, just make a perspective to display completed parents and do your thing.

The benefit is maximal: Nothing falls through the cracks, and everything is kept simple, clear and tidy. No conceptual acrobatics. No mental turbulence.

As an added benefit, this suggestion only requires pruning of the existing code, rather than augmentation.