Is there a reasonable sense in which PA is complete?

Friedman Harvey wrote:

>***Under any reasonable measure of complexity of sentences in the language
>of PA, any sentence or scheme in the language of PA of modest complexity is
>either provable or refutable in PA.***

This assertion surprises me. I for one don't believe it, and see no compelling
evidence for it considering the large number of such sentences that have
remained unresolved despite great interest.

Martin