Re: Docker

From: Steve O'Hara-Smith <steve_at_sohara.org>
Date: Mon, 17 Apr 2023 14:38:24 UTC
On Mon, 17 Apr 2023 08:58:25 -0500
Tim Daneliuk <tundra@tundraware.com> wrote:

> On 4/16/23 23:57, Tomek CEDRO wrote:
> > hat messed up the once beautiful logical and coherent IT world
> 
> With over 40 years in the tech business, I can say with confidence
> that this never actually existed.  The closest we ever came
> to this was the "logical and coherent" world of IBM mainframes.

	I recall one project where used of SSADM was mandated for the
entire project. The "user acceptance" document was a stack of paper about
three feet high, in order to read the thing it was necessary to spread the
pages out five at a time. I was less than half an inch down the document
when I gave up as I was drowning in inconsistencies (I wasn't expected to
sign off on it) three days later we got a sign off on it. That was
waterfall done badly wrong.

	On another connected project we were allowed to go our own way and
do it right, while a team of 'analysts' reverse documented the project as
though we had designed it using SSADM - they had difficulty with some parts
as their tools were insufficiently expressive to describe the solution we
built. That was hacking round the deficiencies of waterfall done badly
wrong.

	There is a great deal to be said for the core principle of agile,
iterative development with constant feedback can be *very* effective at
avoiding many common pitfalls with large projects. Wrapping this simple and
effective principle in a dogma of two week cycles, everybody in the team
can do everything and so forth is less effective and if it is done without
understanding of the core principles and how to use them effectively to
achieve an understood result it becomes an ordered path to chaos. The
quickest way to chaos is to use it to avoid understanding the problem. The
quickest way to success is to use it when you understand enough of the
problem to be useful but not enough to be perfect because nobody has
sufficient imagination. You still have to get the core right!

	The common point where any methodology fails is when those using it
fail to realise that it is a tool for achieving a result and not a goal in
itself. If it gets in the way and doesn't help it's the wrong tool, this is
not cured by getting increasingly doctrinal about it.

-- 
Steve O'Hara-Smith <steve@sohara.org>