View Single Post
  #7  
Old December 23rd 03, 03:33 AM
Pete
external usenet poster
 
Posts: n/a
Default


"phil hunt" wrote in message
. ..
On 22 Dec 2003 17:41:26 GMT, Bertil Jonell

wrote:
In article ,
phil hunt wrote:
On 19 Dec 2003 15:38:09 GMT, Bertil Jonell

wrote:
In article ,
phil hunt wrote:
I've worked as a programmer for
defense contractors (and for other large organisations), and believe
me, there is a *lot* of waste and inefficiency. If the software was
written right, it could probably be done with several orders of
magnitude more efficiency.

What competing method is there except for Open Source?

Open source -- or rather, using some of the ideas from how OSS
projects are btypically run -- is certainly useful.


The reason for my question is that I don't think Open Source is
very applicable the type of 'sharp edge' military systems you are
talking about here.
It is very applicable to making programs that help you make sure
that every regiment gets the correct number of socks and ammo, but not to
making program that handles guidance and target discrimination routines.
Especially not if you expect your capabilities to remain anything
like secret.


Certainly.

Using open source software such as operating systems, IP stacks,
image processing libraries, encryption libraries and the like would
probably be appropriate, and contributing any changes back to those
codebases might well be a good idea. The really secret stuff is much
less likely to be made available.

I also had in mind OSS *techniques*, that is using some of the
procedures in infrastructure that OSS projects often used, to do
closed-source development. Things like Sourceforge, mailing lists,
CVS, packaging as tarballs, etc.


Those are mere techniques to facilitate the actual work. And have little to
do with actually designing a viable weapons system.

The type of paper upon which you compose your missile design has nothing to
do with building a missile.

And a lot of those OSS techniques are not conducive to weapons design.
Folding your mods back into an OSS image processing library, for instance,
is not too wise when you are trying to develop a system in secret. Unless of
course you want your potential targets to know exactly what your system is
looking for (and thereby how to defeat it).

Pete