[SystemSafety] Accuracy of COVID modeling code

David MENTRÉ David.MENTRE at bentobako.org
Fri May 8 21:01:53 CEST 2020


Hello,

Regarding floating-point computation, for once I'll agree with Derek:
accuracy of floating point computations does not depend on the
programming language but the correct use of the FPU. A classical reference:
    What Every Computer Scientist Should Know About Floating-Point
Arithmetic
    https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html

Regarding strong typing:

Le 06/05/2020 à 20:09, Derek M Jones a écrit :
> So called strongly typed languages provide functionality that developers
> have to use to create contexts where checks can be made.
>
> Compare C++ and Ada:
> https://shape-of-code.coding-guidelines.com/2014/04/17/c-vs-ada-which-language-is-more-strongly-typed/
>

In my view, this is precisely the difference between C/C++ kind of
language and Ada and similar: C/C++ does implicit conversion while Ada
requires explicit conversion. I see this has a strong advantage, because
it forces the developer to think about impact of conversion, while
apparently you think C++ implicit conversion is better. And so despite
all the other advantages of Ada your underline in your blog post: range
types, creation of new incompatible types, ...

Best regards,
D. Mentré





More information about the systemsafety mailing list