>   To demonstrate the utility of the conditional construct, consider the
>   following example:

This does not seem like a very compelling example, because it is so
abstract.  So far, it seems that this proposal is sacrificing the
principle of least surprise to something of undemonstrated utility.

I think the proposal would be more interesting if you could exhibit
either a couple of very compelling concrete example or at least a
half-dozen moderately interesting examples, where the utility of letting
the implementation choose is "obvious" and "significant".  In each case
it should be plausible that the implementation can reasonably be
expected to be able to make the right choice.

--lars