yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxxxxxx>]

On 11/15/18 5:16 PM, Jerome Duriez wrote:
> I'd thus propose to get rid of ZERO_TOLERANCE at least, replacing with 
> EPSILON where it is currently used.
Is it worth it?...

> (As for me, tolerance would for instance be  useful to know when an 
> increasing value (x, increasing from x0 < a) can be detected as 
> entering an interval [a,b], accounting for a possibly blurred 
> comparison between x and a. Would you consider this as very specific ?
Why would you say that a value lies in an interval when it does not? I 
can't imagine any realistic situation like this.
Detecting duplicates in stl is maybe a real specific situation (because 
writing numbers in ascii leads to roundoff errors), but I'm not sure 
even for that case.

Bruno