Migrated from my old blog from 2015.
“Agile is just some new fad” might have been a reasonable objection in 2001 but it seems less reasonable after more than two decades.
Let’s look a little closer at the history…
When all the lightweight methodology people got together at Snowbird in 2001, Extreme Programming (aka XP) was by far, the most popular approach. XP was born when Kent Beck was asked to rescue the C3 project, a payroll system replacement at Chrysler, in 1996. I first heard of XP in late 1998 on comp.object, comp.software-eng (back when Usenet was still a thing) which led me to Ward Cunningham’s Portland Pattern Repository (aka the first wiki) and Ron Jeffries’ xprogramming.com (imagine current version with late 1990s web aesthetics). So Extreme Programming has been around for close to 30 years.
The Portland Pattern Repository has a pretty good summary of the timeline.
There is more in the History of Extreme Programming. For example, the emphasis on unit testing (not yet test-first) came before C3. And you can see a lot of ideas leading to Extreme Programming in Ward’s Episodes pattern language of competitive development submitted to the Pattern Languages of Programming conference in 1995.
Now one might argue that Extreme Programming is no longer the face of Agile but take a look at slide 22 on one of Eric Ries’ early presentations on Lean Startup in 2008:
The current face of Agile, which I’d argue was primarily caused by certifications, is really Scrum. Scrum started before XP in the early 1990s; XP borrows structural elements from Scrum. I am less familiar with Scrum origins but from what I understand it was derived from multiple projects in the early 1990s, one of which was developing Quattro Pro at Borland in 1993. The more official description, and I think the first mention of “Scrum”, was presented at OOPSLA in 1996. So Scrum has been around for around 30 years.
If you look even closer, it gets yet more complicated.
Ward and Kent worked together at Tek Labs, formerly the R&D labs at Tektronix. I’m guessing this was in the 1980s given their paper introducing Class-Responsibility-Collaborator (CRC) cards was presented at OOPSLA 1989, and by then they were both working elsewhere. The claim is that they rediscovered pair programming when they worked at Tek Labs, so XP-style pair programming has been around for probably over 35 years.
In 2003, Craig Larman and Victor Basili wrote an article for IEEE Computer about the history of iterative and incremental development which walks the path starting from Walter Shewart’s emphasis on short PDSA cycles in the 1930s to modern Agile methods. The evolution of development towards iterative and incremental has been happening for over 80 years — Agile is simply the latest expression of it.
