[graap-wg] OneOrMore / ExactlyOne

[Constantinos A. Kotsokalis]

Hi all,

A question regarding the two term compositors mentioned in the subject:
Should one expect that ExactlyOne is the non-greedy version of OneOrMore?
That is, should an implementation attempt to satisfy as many terms of a
OneOrMore compositor, whereas for an ExactlyOne compositor it would just
quit as soon as there is found a term which is successfully evaluated
/ executed?

In both cases, should it be a requirement for all terms to be valid
against the agreement constraints, even if there are terms who satisfy
them (the constraints) and were successfully evaluated / executed?

Or are the above completely implementation-specific?

Best regards,

  Costas