Why computers calculate wrong

Why computers calculate wrong

By Karl Brodowsky
Datum: Samstag September 3, 2016 14:10
Dauer: 40 Minuten
Target audience: Jedes
Language: English
Tags: arithmetic perl5 perl6

Mehr Informationen gibt es auf der Seite des/der Vortragenden:


We have in Perl 5 and Perl 6 the first challenge to know and control which numeric type we are using. Floating point arithmetic is inprecise and I will show examples how it can fail.
Limited size integer arithmetic is often performed modulo a power of 2, so it will roll over silently. When this behavior is not expected and anticipated, it can cause subtle and hard to understand errors that occur much later than the deployment of the software. Rational as numeric type has the weakness of using a lot of memory and time when doing more complex calculations.
But with both Perl 5 and Perl 6 we have good ways to handle these issues. But it is off course applicable to most programming languages.
I have blogged about issues related to this a lot, see my blog in the URL


Attended by: Markus Hechenberger (‎Hechi‎), Michael Kröll (‎pepl‎), Lee Johnson, Renee Bäcker (‎reneeb‎), Dirk Deimeke, brian d foy (‎brian d foy‎), Jozef Kutej (‎jozef‎), Wolfgang Schemmel (‎Perleone‎), David Schmidt (‎davewood‎), Stefan Hornburg (‎Racke‎),