Re: Proposal: Add log1p and expm1 to GHC.Float.Floating

The problem there is that now you can't just replace log (1 + x) with log1p in your code, you have to reason through case by case about an open universe of instances you don't own, to know whether it is a safe transformation. I use type classes to quotient that sort of case by case reasoning out of my thinking.

On Apr 20, 2014, at 9:25 PM, wren romano wrote:

...

On Sun, Apr 20, 2014 at 8:49 PM, John Lato wrote:

...

On Sun, Apr 20, 2014 at 5:20 PM, Edward Kmett wrote: I know for me personally it would force me to double the amount of numeric code I write, just to maximize my audience. I really don't want to go there. I just want to be able to call the function I mean, and to be able to talk to the right people to make it do the right thing.

exp1m = error "Go bug some library author to implement exp1m"

would accomplish that even more efficiently, since it will directly point users to the right people.

FWIW, because log1p/expm1 etc are not part of the current API, this too is a strictly monotonic improvement over the current state of things: users who don't know/care about the new functions don't need to change anything, and users who do care are beholden to see that they are implemented correctly. No need for CPP. And no misleading implementations of the fused functions.

I'm +1 for these defaults.

-- Live well, ~wren _______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries