Re: WSJ article on software liability

From: Traveler (traveler_at_nospam.com)
Date: 02/28/05 

Date: Mon, 28 Feb 2005 08:42:19 -0500

In article <112663uh51rsb94@corp.supernews.com>, CTips
<ctips@bestweb.net> wrote:

[cut]
>Get over this idea that hardware design is comparable in complexity to
>software programming. The largest hardware designs are nowhere close in
>complexity to even large software projects - and we won't even talk
>about things like ATC programs or fly-by-wire avionics suites.

My point is not that one is generally less or more complex than the
other. My point is that an algorithmic software system of a given
complexity is much more unstable than an equally complex hardware
system. The hardware systems used in my computers have never failed
even though the interactions between the various chips are very
complex. They don't fail for mainly two reasons: 1) the timing of the
signals is deterministic and 2) the event dependency problem that
plagues software does not exist in hardware.

>BTW - any chance we could see an implemenation of a simple sort routine
>for 1e8 numbers in your favorite COSA language? I'm still waiting to see
>if you can come up with anything that doesn't look very similar to a
>sort routine written in a mainstream programming language.

I realize that what you are looking for may be important to a lot of
people but why should a COSA routine not look like a routine written
in another language? COSA does not prohibit the use of sequential
steps for problem solving. One needs to understand the difference
between an algorithmic and a non-algorithmic system. In the former,
only two elements (instructions or operations) can communicate at any
one time (a predecessor and a successor). In a non-algorithmic system,
by contrast, the number of elements that can communicate at the same
time is indefinite. See the section on the Silver Bullet page about
"Programs as Communication Systems" to get a better idea of what I am
talking about.

The primary reason for using a signal-based, synchronous software
model is that it provides us with *the* solution to the biggest
problem plaguing algorithmic systems: unresolved event dependencies
otherwise known as the blind code problem.

Louis Savain

The Silver Bullet: Why Software Is Bad and What We Can Do to Fix it
http://users.adelphia.net/~lilavois/Cosas/Reliability.htm

Relevant Pages

  • Re: WSJ article on software liability
    ... >) hardware is simpler than that of software. ... >) hardware chips interacting in highly complex ways and yet they have ... of behavioral complexity, almost all hardware systems remain logically ...
    (comp.programming)
  • Re: WSJ article on software liability
    ... >)>size and complexity of a software emulator of a chip. ... Does that make hardware any less complex? ... A software emulation of hardware is a lot more stable ... Hardware failure is so much more expensive than software failure. ...
    (comp.programming)
  • Re: OT: CPUs
    ... strategy but rather yields oscillating trends: (complexity in hardware ... There certainly are numerous examples of coevolution in predator-prey ...
    (talk.origins)
  • Re: WSJ article on software liability
    ... >>'' My point is that an algorithmic software system of a given complexity ... >> is much more unstable than an equally complex hardware system. ... >>are either using the assumption that they are of equal complexity, ... hardware design vs. software design. ...
    (comp.programming)
  • Re: A quote from Crypto-Gram
    ... >> programming modes to controlled by a hardware input. ... Controlling ... > hacker/virus maker to begin altering the microcode. ...
    (sci.crypt)
Loading