I like the terms used here of 'deterministic' and 'partially deterministic' and concur that not only do we see the 'variable reality' as perceived by the end-user of computerized systems we develop, but that those systems can actually alter that reality...

I agree with you that theoretically there may be some "highly algorithmic and well defined task" where a H2O-Fall may work as well as an agile approach. I don't know if I have ever actually participated in such a project, but I suppose it is possible. Nevertheless, I can't see the downside to using an agile, iterative approach even on these types of projects. What is the down side for delivering the highest priority portions of that system earliest? What is the downside to getting customer eyes on the product earlier to make sure your "highly algorithmic and well-defined task" was not misinterpreted by the coders. Give me a good agile iterative approach any day. Small deliverables early and often, inspection and adaptation, etc

Very good treatise that on System influences reality. I have been ranting same here on system altering the user behavior. I wish this concept gets mainstream and gets tackled better.

Now back to Waterfall model, I do not believe there is any problem in trying to determine as much as you can before embarking on design and development. The problem is trying to implement everything in one go. This takes away our ability to learn from mistakes and apply corrective actions.

“…I can't see the downside to using an agile…”, one of my readers said. This is one of the reasons why I wrote the article in the first place.

As we know, Agile is a mindset. It cannot be reduced to a number of techniques. However, we often hear: “Use X, and it’ll solve all of your problems”. No, it won’t! “You cannot be agile unless you do Y and Z”. Yes, you can! When an approach that is based on common sense is reduced to a set of religious beliefs, it loses common sense. And then everyone loses – us and our clients.

When we get excited by something – a new book, a new friend, or a new software development methodology – it’s not easy to step back and look at our new “acquisition” with a critical eye. Yet, everything has its limitations, its boundaries. Remember, Newton’s physics didn’t die with Einstein’s breakthrough discovery. But it’s boundaries of applicability were redefined. We still happily use it every day without thinking about
relativistic effects.

What are the boundaries of applicability for a development process, be it waterfall or Agile? “Give me a good agile iterative approach any day” may be an expensive proposition. Every release requires regression testing. Auto-tests do help, but in their turn require maintenance, ever more with every new test. Add to it preparation and conduct of a promotion (often risky), potential disruption of operations and user re-training. All in all, I can see how Agile approach can become more expensive than Waterfall for straightforward algorithmic tasks with stable requirements.

As I said in my article, I have not seen this kind of projects yet during my 30+ years in software development. But that doesn’t mean they do not exist. After all, I only worked in the enterprise applications field.

It would be a stretch to paint me as an adept of Waterfall. As a member of Vancouver Agile council, and a leader of a team that practiced Agile for a few years with impressive results, I am sold on Agile. One can see my previous posts in this blog: “Agile Software Development - A Cinderella Story?”, “How One CIO was Converted to Agile”, “A Client as a Teammate” and others. However, I see Agile becoming more and more of a religion, and this is a very unfortunate development.

Regards,

Eugene Nizker

PS. A friend of mine who taught me how to write meaningful code some 30 years ago, Alex Zakharyonok, commented on another thought in my article: the systems we develop influence the reality. Alex says: “For business applications, I would formulate it more aggressively: we should only develop systems that change the reality; otherwise they aren’t worth our time”. I can’t agree more.