Re: Critique of SEI's Ada DARK project

[spray@convex.com (Rob Spray)]

One concern I have always had about DARK is how it got funded
targetted to a processor that is not available in a 
radiation-hardened configuration.  Several embedded systems
I am familiar with are required to use full MIL-SPEC processors.
My experience with several distributed systems is that despite all
good intentions, they are very processor dependent.  I suspect
that DARK would have been much harder on a processor like the
Intel 80286 or 1750A that were the rad-hard choices that had 
an Ada compiler when DARK was started.

Now, in that time-frame, the state-of-the-art for compilers 
and runtime systems for the 68K was far ahead of the other 
machines.  Also, workstations like the SUN and Apollo were 
available for hosts and targets.  So the costs and risks 
to DARK's preceived success were lower with the 68K.
However, the result is a system that, IMHO, has minimal
chance of being ported to a rad-hard environment.

Tell me I'm wrong.

--Rob Spray
--spray@convex.com
--214/497-4110

Disclaimer: This is my own cynical conjecture, and may not
	    reflect the opinions of employers past or present.
	    Your opinions may vary. Ted Holden's opinion will
	    doubtless be different. But then I would bet I've 
	    worked on more rad-hard distributed Ada systems than
	    he has.