[Info-vax] DCL's flaws (both scripting and UI)

[John Reagan]

On Saturday, January 31, 2015 at 12:57:29 PM UTC-5, glen herrmannsfeldt wrote:

> I think all processors can reliably detect a floating point zero, and
> most reliably detect equality on small integers. Those happen often
> enough that I wouldn't want to disallow them. 
>  

Pos zero and neg zero matter to you?

> 
> A high-level language with a convenient "approximately equal"
> operator might be nice.  You might want a way to define the delta.
> 
Ada has T'MODEL_EPSILON, Pascal has EPSREAL, Fortran has EPSILON() to help you in this area.  It isn't the same as "almost equal", but you can do something like

ABS(A-B) <= EPSILON 

or such.