Printing floating points

[Ola Fosheim Grøstad]

On Friday, 29 January 2021 at 02:58:38 UTC, Bruce Carneal wrote:
> The parenthetical phrase was meant to apply to everything that 
> preceeded it.  An SMT like solution is what I thought you had 
> asserted, something that substantially cut down the work while 
> provably covering the full domain.

I am not asserting anything, other than that I don't think full 
formal verification (unless bitblasting works) is within the 
perimeter of the D community, so that leaves informal proofs 
combined with selective testing as the viable option that the D 
community as a whole (not any specific individual) can accomplish.

A solver is a tool that can be used to make this viable. Without 
a solver, maybe a reimplementation with a correspondence proof is 
easier.


> better than continuing here.  I'll be hanging out at beerconf 
> if you think it's worth your time to converge on understanding.


I like the idea of beerconf, but I have a deliverable deadline on 
monday, so I don't have time this week.