Friday, February 17, 2006

Falling over the waterfall

I'm amazed by how many people I meet that tell me their project, or projects in their organisation, use some kind of variant of the waterfall lifecycle. It happened to me again today while conversing with a friend who works for one of the major international banks. The waterfall is as ubiquitous as death, taxes and project overruns!

[And now I discover there's an unmissable conference on the subject - stop reading this blog and go now to! But more on this later.]

The interesting thing about the waterfall lifecycle is that it has very few proponents among the luminaries of the industry. Even Winston Royce, who is usually credited with inventing the waterfall in his 1970 paper to the IEEE was actually criticising the approach of trying to deliver a complete system in one iteration. While I'm in name-dropping mode, I could mention that I met Winston when he was working at TRW. He had a most distinguished pedigree in software engineering but he was hardly the unqualified supporter of monolithic processes. His son follows in his tradition and provided a very interesting quote on the waterfall process:
  • "Across the software industry, we characterize modern software lifecycles using many different terms, including spiral development, incremental development, evolutionary development, and iterative development (my preference). In spirit, these terms all stand for the same thing, namely anti-waterfall development."” Walker Royce (2000)

More striking criticism of the waterfall approach comes from all quarters. F. P. Brooks for example:
  • "Much of present-day software acquisition procedure rests upon the assumption that one can specify a satisfactory system in advance, get bids for its construction, have it built, and install it. I think this assumption is fundamentally wrong, and that many software acquisition problems spring from that fallacy.”" F. P. Brooks (1986).

Or Tom Gilb:
  • "There is nothing (no complex thing) that can't be delivered in an evolutionary fashion; conversely no (complex) thing can be delivered in one go." Tom Gilb.

The waterfall has a significant advantage. It's simple to explain and understand. For those tempted to adopt it for that reason should heed this warning from the celebrated journalist, political commentator [and cynic], Henry Mencken:
  • "For every complex problem, there is a solution that is simple, neat, and wrong!"” H. L. Mencken “.

However I would want you to think that I'm biased in any way. I recently was pointed to this new conference on "Waterfall Unified Process" by a friend and former colleague Jon Kern. Whatever your view on waterfalls, I urge you to visit this site. It's great!

I've already mentioned the link. It's
Post a Comment