[AusNOG] Looking for employees? Melbourne based

Paul Jones paul at pauljones.id.au
Fri Oct 11 19:18:52 EST 2013


They still teach it at USQ as part of comp sci and engineering. Reasoning was we need to learn how floating point works so we can determine what sort of accuracy our application requires.
We had to do something like 0.423e-3 -12.234e2 in the exam. Mantissa and exponent ring a bell? :-)
Also had to do binary, hex and octal manually. Sounds hard but it's actually quite simple to do.


Paul.
#2

-----Original Message-----
From: AusNOG [mailto:ausnog-bounces at lists.ausnog.net] On Behalf Of Paul Brooks
Sent: Friday, 11 October 2013 6:48 PM
To: ausnog at lists.ausnog.net
Subject: Re: [AusNOG] Looking for employees? Melbourne based

On 11/10/2013 1:31 PM, Phillip Grasso wrote:
>
> 0.2+0.1 = 0.30000000000000004
>
>
I vaguely remember hand-calculating binary approximations of floating point arithmetic with pencil and paper, with a few left and right register shifts for scaling thrown in, to demonstrate these issues during CompSci at university. They actually taught it as a module.

Probably to make us aware of the dangers (or opportunities) of leaving computers to do floating point arithmetic with large numbers, allowing white-collar crims (or
opportunists) to make a fortune off the missing rounding bit-errors.

'Course, that was a while ago now....

(One of my favourite employment filters for network ops:

How long will a gigabit ethernet router port at full utilisation cause the byte-counter to wrap, if the byte-counter is a 96-bit register)

Paul.

_______________________________________________
AusNOG mailing list
AusNOG at lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog



More information about the AusNOG mailing list