Subject: *** PROBABLY SPAM *** Re: rocket science
From: Laurent GUERBY <laurent@guerby.net>
Date: Sun, 27 Feb 2005 20:06:53 +0100

On Fri, 2005-02-25 at 22:05 -0500, Robin 'Roblimo' Miller wrote:
> Note that the shuttle onboard computer group works without
> competition. There is no "Joe's Shuttle Software" down the street
> selling a competing  product. There is no pressure to produce new 
> features in order to maintain or increase market share.
>
> In other words, this vaunted software project is the antithesis of 
> private enterprise, freedom, and corporate efficiency in almost every 
> way, right down to stifling creativity. It has more in common with 
> Social Security than with the Mplayer project.
> 
> In return, we see nearly error-free software. And while this is nice,
> we need to remember a basic bureacratic dictum: that the best way to
> avoid mistakes is to do nothing.

Note that nowadays rocket code (eg: EADS/Ariane) is written for
commercial purposes and competing in a global market. If you're serious
about business, you'll just look at the cost of failure and
market forces will make you spend much more on your flight software than
on say what MPlayer would cost.

EADS tried cost cutting by omitting some software testing and by
attacking the "expensive" software culture, they won't do that mistake
ever again, it would be suicide in the launch market.

I don't know the numbers, but I wouldn't bet so lightly like you did
that NASA shuttle software development is uncompetitive price wise (if
the NASA process would be reused for software in the satellite launch
"rocket science" market of course).

And "do nothing" is not an option in the satellite launch market.

Laurent