From: julio@diegidio.name   
      
   On Tuesday, 29 November 2022 at 19:02:19 UTC+1, Dmitry A. Kazakov wrote:   
   > On 2022-11-29 18:37, Ben Bacarisse wrote:   
   >   
   > > In my view, if you need to check the arguments, that's a separate   
   > > function because what to do with bad data is so often   
   > > application-specific.   
   >   
   > Yes, validation is a part of the measurement process. Usually datatype   
   > in which sensors report data and ones used in computations are   
   > different. Validation happens upon conversion and then, ideally, invalid   
   > values are excluded per design.   
   >   
   > > Just normalising and carrying on is only one   
   > > strategy, and one that can delay detecting bugs.   
   >   
   > Sure. E.g. IEEE 754 float design is a perfect example of such bugs when   
   > computations with non-numbers just continue until too late.   
      
   A couple of very nonsensical statements.   
      
   But congratulations especially for always   
   fighting the good fight...   
      
   Julio   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)