I appreciate your original suggestion and your bringing the focus to bear on 'first principles'. I understand and support the principle of cohesion.

We may differ on the application of that principle to this case. I submit that "give me a value or tell me why you failed" is sufficiently cohesive and a useful job description for a subroutine. Granted, it is not Functional Cohesion in the strictest sense. But I think it is a common, intuitive, and effective way to foster clear program flow at the level of the call.

I agree with your comment on btrott's submission and have /msg'ed him to say so. Since my app is already OO, it is the approach I intend to use in this case.

But not every program is OO and I was simply wanting to build on your very helpful suggestion and see if it might be possible to extend that idea to allow a useful return value while factoring out the wrapper subroutine.

My question (and it's about time for me to stop yacking and run some test scripts) is whether the   eval {die "error description"}   trick will work and allow me to access $@ back at the call level.

Thanks again for this exchange. David