Prev: Re: Off Topic: Date Stamp (was: Another fighter question) Next: DSII Question

Re: [OFFICIAL] Some FT background stuff (guidelines for writers)

From: Thomas Barclay <Thomas.Barclay@s...>
Date: Sun, 15 Feb 1998 21:06:09 -0500
Subject: Re: [OFFICIAL] Some FT background stuff (guidelines for writers)

> >I'm a programmer, and I will say that programming is still more art
> >science.  We can't take a measure of a program the way we take a
> >of a bridge or a starship and say it's going to work or not.  
> I'm also a programmer (as well as systems analyst and technical
> writer) and we can do a much better job than we are doing. I do know
> what you are saying. It's the unknowns that kill programs. You can
> overcompensate with a bridge by building it to 3 times it's necessary
> limits. It's hard to do that with a computer. However, it can be done.
> NASA does it by having three versions of the computer/software onboard
> the shuttle, and requiring two of the software versions be written by
> two different teams independently. It can be done.

1. We can do better than we are doing. Probably explains why 
traditional engineers frown on programmers. (Of course, the 
programming model differs from the engineering model - you don't 
release a bridge that might GPF and fix it in the next rev....)
2. To build a bridge 3x thicker might cost 1.3x as much. To build two 
sets of the same program fully independently might cost 2.0x as much. 
So I don't think you can quite compare the two cost wise. 
3. Program correctness has long been a discipline studied with the 
goal of making a language where one cannot author bugs easily or at 
all. It just turns out that (according to graphs I've seen), as 
correctness increases, usability decreases. And in any extremely 
complex system, you can unit test each subfunction, and then test 
chains of functions, but what is really hard to test is the operation 
of the entire software product as it has a lot of unexpected 
potentiations. Also note that we only ever test for things we 
expect.... which leaves a huge gap sometimes....

(In short, its a tough problem, and one the consumer often does not 
want to pay to solve - Joe Consumer says "I want good software" and 
then goes out and buys the latest Windoze release KNOWING it will be 
buggy - "Gotta have that new XYZ Technobable technology thing...." - 
wouldn't do that with a car I hope!). 

Thomas Barclay
Software Specialist
Police Communications Systems
Software Kinetics Ltd.
66 Iber Road, Stittsville
Ontario, Canada, K2S 1E7
Reception: (613) 831-0888
PBX: (613) 831-2018
My Extension: 2034
Fax: (613) 831-8255
Our Web Page:

Prev: Re: Off Topic: Date Stamp (was: Another fighter question) Next: DSII Question