The main problem in software engineering is figuring out what is this
program supposed to do, and how do we verify it?
The Orion (only to pick one example) has an extremely complicated
interface to the user (possible combinations of parameters and
sequencing), which is probably not written down or specified anywhere.
This is in contrast to the RF specs, which are fairly easy to state
and measure.
Most software design is seat of the pants - you'll know it when you
see it, etc. Formal definition of software specs and correctness is
something you see at NASA and Boeing (well, sometimes!).
I am afraid that K4TAX and others are right that this is the way of
the future. The boat anchors didn't have such great performance, but
at least they were predictable, and you could figure out most problems
from a schematic. (I keep a TS-520S in the closet, in case of EMP
attack or other digital meltdown.)
73 Martin AA6E
On 12/19/05, ChasW3KC <w3kc@verizon.net> wrote:
> Yes, bugs can be lurking in complex software. When I was in
> the real time software business, our idealistic goal was to
> have the system fail gracefully - or at a minimum, allow a
> dump of critical info to aid in debugging.
> 73 Chas W3KC
>
> _______________________________________________
> TenTec mailing list
> TenTec@contesting.com
> http://lists.contesting.com/mailman/listinfo/tentec
>
--
martin.ewing@gmail.com
http://blog.aa6e.net
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|