[PRL] proper tail recursion

[Dave Herman]

> I don't see the connection.  Can you clarify what you're thinking, Dave?

Ah, probably not -- I asked the question before having thought it 
through carefully (sorry).

I'm reading some papers on linearity, and they keep talking about how 
when you can prove that a resource is only used once you can implement 
it with destructive update, and this reminds me of proper tail 
recursion: when you can tell that the control context will not be needed 
again, you can destructively replace it. I wondered if there's a way of 
talking about control in terms of linearity.

Dave