I updated the wiki page on Waterfall today, nearly doubling its length to what is now an article roughly 35 pages in length if printed. As with the previous version, I’m trying less to describe the common interpretation of what ‘Waterfall’ is, and more trying to show that the common interpretation seems to be wrong based on the historical documents. Or at least that the common interpretation of ‘Waterfall’ may never have existed as anything more than a straw man.
I’m not a big promoter of “Waterfall”, although I am a big believer in analysis and design up front (both of which can evolve later, but the more you do early the better you understand your current situation and where you want to go). But the thing I find fascinating about ‘Waterfall’ as a concept is how like a mystery novel it is.
Poor Winston Royce wrote what at the time was a non-controversial paper that discussed current practices for large software development and in which he made some relatively minor suggestions for improving things. He doesn’t use the term ‘Waterfall’ anywhere. Yet starting almost a decade after he published his paper, and continuing to this day, his name is tied to something called ‘Waterfall’ that bears very little resemblance to what he wrote and which is called such things as a “toxic concept” and “the most costly mistake in the history of the world”.
Yet to this day I can’t find any document that EVER advocated for anything close to the highly rigid process that is how ‘Waterfall’ is commonly described. Describe such as process? Sure. Usually while citing Royce’s paper. But not advocated.
The new article is long, hopefully informative, fully referenced, and spends way more time talking about history than it probably should. Be aware that I have no outside editor, so there are probably grammar and other mistakes galore.
Feel free to add comments if you have any.