[Discuss] Software Development (OT?)

[Jerry Feldman]

On 09/09/2013 08:57 PM, Richard Pieri wrote:
> I suggest that software development hasn't really changed per se. It's 
> always had a strong focus on processes. Even when I was hacking little 
> things some 30-odd years ago I was following processes. They hadn't 
> been formalized yet but the processes were there. What's changed is 
> scale.
>
> If you were writing code for a living 30-odd years ago you were 
> probably responsible for entire programs from start to finish. Maybe 
> you worked with a partner. It didn't matter if the programs were 
> simple loan calculators or office suites or operating systems. It 
> didn't matter if you were a one-man shop or you worked for a giant 
> like Xerox. You could use whatever procedures were most comfortable 
> for you, change them on a whim if you wanted, as long as you delivered 
> your programs on time and on budget.
>
> If you write code for a living today then you're probably part of a 
> group, an organization in fact, and possibly one that's part of an 
> even larger development organization. You're not really a programmer 
> in the old sense. You're an assembly line worker. You assemble the 
> pieces of code that your managers tell you to write. You follow the 
> procedures set down by those managers because if you don't then your 
> code won't work with the pieces of code being written by all the other 
> workers in the organization. You follow the procedures because you'll 
> be fired if you don't and you'll be fired if you complain about them.
>
I've been writing code form almost 50 years (with a 3 year break flying 
helicopters in Viet Nam). While back in the 60s I just knew Fortran and 
BASIC, but in the 70s when I worked for a large company, some of the 
software we purchased and customized, but there has always been a 
philosophy of design-code-testand evidence that better code was produced 
if it were designedand the design was properly reviewed, and that the 
guys who wrote the code participated and then followed the design. The 
main issue here is to find any problems in the design. Time frame here 
is the late 1970s. Much of my time after that was working with compilers 
and run-time libraries and porting code between architectures.

-- 
Jerry Feldman <gaf at blu.org>
Boston Linux and Unix
PGP key id:3BC1EB90
PGP Key fingerprint: 49E2 C52A FC5A A31F 8D66  C0AF 7CEA 30FC 3BC1 EB90