Re: Towards better embedded software (long

From: Steve at fivetrees (steve_at_NOSPAMTAfivetrees.com)
Date: 05/03/04 

Date: Mon, 3 May 2004 00:24:23 +0100

"CodeSprite" <debate@codesprite.com> wrote in message
news:Xns94DDB8EEB8384oHcRiKeYwotID@63.218.45.215...
>
> Well I don't know... my attitude is "all software is guilty until proven
> innocent" - also known as debugging by bloody-mindedness. There again, if
> I've eliminated errors before releasing code, couldn't I say there were no
> bugs and therefore I hadn't been debugging? I wouldn't mind too much if
> someone said he was good at debugging, if it meant he actively looked for
> problems in his own or other people's code, rather than glazing over and
> saying "looks ok to me".

Point taken, but I talk in terms of "good design" and "verification" rather
than "debugging" ;).

> One problem is that we call them "bugs" - makes them sound cute and
> whimsical. They are MISTAKES caused by failing to think through the
> problem, and the longer they go undetected the more expensive they are to
> fix.

Exactly. I can't think of any design activity that I'm involved in where
failure is accepted so casually and routinely - other than software.

Steve
http://www.sfdesign.co.uk
http://www.fivetrees.com

Relevant Pages

  • Re: Towards better embedded software (long
    ... >> I don't believe in debugging, ... >> encouraging this attitude, but engineers are the ones who embody it. ... I've eliminated errors before releasing code, couldn't I say there were no ... bugs and therefore I hadn't been debugging? ...
    (comp.arch.embedded)
  • Re: Debugger features
    ... the problem was at each stage, without debugging. ... had shallow bugs, within the reach of printf and my IO ... However macros ... the interaction between the control structure and the code ...
    (comp.lang.lisp)
  • Re: TDD doesnt suck
    ... so maybe JUnit is big because Java is a superior language. ... adding nothing - simply left Java open to grow a whole new category of bugs. ... Under TDD, you run all tests after the fewest possible edits. ... Teams using TDD overwhelmingly report much less time spent debugging. ...
    (comp.lang.c)
  • Re: [patch 00/13] Syslets, "Threadlets", generic AIO support, v3
    ... bugs showed up as a subtle epoll performance ... And debugging state machine code has exactly the same complexity as ... with, even in the most simple, TCP socket case! ... data is already on the kernel stack. ...
    (Linux-Kernel)
  • Re: TDD and the development time
    ... That means if something fails unexpectedly, ... checkpoint (and the risk that checkpoint also had bugs). ... debug deep into some awkward corner, change the code, reposition the control ... flow point, recompile, and resume debugging without stopping the run. ...
    (comp.object)