It's pointless to speak of usefulness without specifying a utility function.

It is just as possible to differentiate as it is to integrate.

If it is determined a priori that unfalsifiable propositions are not useful, then knowing the result of the Halting Problem is not useful. Isn't that silly?

I strongly object to categorizations which discriminate against valid science (knowledge? truth? understanding? reasoning? Useful facts?). Is all.

The human process of trying to udnerstand reality is continuous, not discrete, so it's silly to reason about it in terms of discrete categories. It necessarily leads to confusion; and the sort of gatekeeping and self-justification Carl Sagan is guilty of.

Science benefits much more from being defined too broadly; than being defined too narrowly.

I'd rather be too permissive then ignore the junk; than be too restrictive and never even encounter good ideas which were erroneously discarded as junk.