> The requirements for the SDI system are, according to the estmates for
> the numbers of lines of code, fully within today's capability, as
> every major operating sytem runs with more lines of code today. What
> is illuminating is that every time industry capabilities reach the
> alleged amount of code SDI's detractors claim is necessary, they move
> the mark.
Redell's argument isn't that the system is too big. He's worked on big
systems. His argument is that existing big systems that work are used and
tested in non-critical situations before they're deployed in the real
world. As they get more use, people are willing to rely on them more.
When a large reservation system, or a new telephone switch, or the control
system for a airframe is developed, it can be tested in simulations, and
then with larger and larger groups of testers. Each change in the deployed
environment reveals more bugs, which are worked around, lived with, or
repaired. Each change in the environment is accompanied by warnings to the
users to expect glitches. SDI isn't going to be deployed to protect small
towns from small barrages, then larger areas from larger attacks, etc. It
will be tested in larger and larger simulations, but in the first attack,
it has to perform flawlessly, or it won't have made things much better.
> Which is why deployment is being deferred for more research with other
I hope that's the case. That isn't the way I remember Bush's rhetoric.
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:39:44 MDT