Re: [Haskell-cafe] Proof that Haskell is RT

On Wed, Nov 12, 2008 at 12:35 PM, Lennart Augustsson wrote:

Actually, unsafeInterleaveIO is perfectly fine from a RT point of view. Since you are talking to the outside world any behaviour is acceptable. All the weird interactions between getContents and writing the same file from the same program could, in principle, happen if a different program wrote the file.

Yes, I was thinking about this on my way to work and thought that I may have spoken too soon; I couldn't come up with a way for unsafeInterleaveIO to break referential transparency (although I couldn't prove to myself that it couldn't, either). Your argument seems good to me, though. -- ryan