G.Gonzales, Comonad Challenge Exercise

[sassa_nf]

http://www.haskellforall.com/2013/02/you-could-have-invented-comonads.html
Challenge Exercise: Prove that our original implementation of up' does not play nice with extend. In other words, find a counter-example that proves the following equation false:

extend (\t' -> p (up' t')) t = extend p (up' t)

What does he mean?

1. There is no need to look for counter-example, since it is false for any t, if we were to compare literally.

2. His previous statements about up' vs extend up were based on behavioural equality of the two, so motivating these challenges to find out which one was "correct". The behaviour of both sides is identical for all t. No? I tried even some cleverly placed undefined.

Comments

[zeit-raffer.livejournal.com]

'undefined' and 'seq' always break things

[sassa-nf.livejournal.com]

Yes, but in that article Mr. Gonzales is looking at which of the implementations is "better". Ie in the context it would mean one of the implementations "behaves well" with extend, the other one doesn't, and this equation is seen as the decider.