From: tkoenig@netcologne.de   
      
   mark horn  schrieb:   
      
   > [[Mod. note -- I think "floating-point rounding errors" is the phrase   
   > you're looking for.  If v/c is very close to 1, then the formula for   
   > gamma tends to be very sensitive to rounding errors, causing the sorts   
   > of anomolous behavior you noticed.   
      
   I would suggest that everybody who does floating-point calculations   
   should read the famous "Goldberg paper", "What Every Computer   
   Scientist Should Know About Floating-Point Arithmetic" available   
   from https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html   
   (it is not absolutely necessary to follow the proofs as a user).   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)