Most years, just after the first of January, I fly to Hawaii and while escaping the New England winter is a bonus, my real reason is to attend the Hawaii International Conference on Systems Sciences (HICSS.) The conference is pretty broad, but one of the tracks focuses on research into Agile practices. Most years I present a paper, this year I shared some of the latest research I’ve been doing on how teams that finish early accelerate faster. Conference attendees voted it the best paper in the Agile track and nominated it for best paper overall.
A paper that grabbed my attention with its title alone is: State-of-the-Art: A Systematic Literature Review on Agile Information Systems Development. (Sorry no link yet, I’ll post it when it is made available online.) In it, Markus Hummel of Goethe University in Frankfurt takes a broad look at the entire state of play in Agile research over the past year. He delves into how many papers were published, who authored them, what subjects were covered, what research methods were used, etc. His conclusions were inconclusive, mainly because he points out a lack of academic rigor in many papers. Hummel calls the field “nascent” and concludes, “Findings of the literature are mostly based on experiences which lack empirical and theoretical support.”
But what really jumped out at me was this comment:
The definition of agility remains one of the most salient problems of agile [Information Systems Development]. A universal understanding of what constitutes ‘agility’ is not observable. Attempts of research-based taxonomies for pinpointing the concept of agility build the basis for a common definition, but most papers still rely on the Agile Manifesto that consists of unverified principles and practices of practitioners, which are not suitable as a solid theoretical grounding. The use of research-based definitions of agility should be extended in order to enable a better comparison of studies and to increase the value of research on agile ISD.
Basically, he’s saying it’s hard to assess the state of Agile development methods because what it means to be Agile has never been defined! Only 45% of the papers even referred to a definition of Agile, most often the Agile Manifesto.
(Scrum was developed totally by a data driven approach and benchmarked by productivity analysis tooling at Capers Jones company, Software Productivity Research in 1993. Only when it performed 10 times better than waterfall was it released. Then we were certain that even a reasonable implementation of Scrum would double performance. It looks like I need to write a paper on this so the academics will understand better why Scrum based experience influenced the writing of the Agile Manifesto.)
(Scrum was developed totally by a data driven approach and benchmarked by productivity analysis tooling at Capers Jones company, Software Productivity Research in 1993. Only when it performed 10 times better than waterfall was it released. Then we were certain that even a reasonable implementation of Scrum would double performance. It looks like I need to write a paper on this so the academics will understand better why Scrum based experience influenced the writing of the Agile Manifesto.)
Agility by its very nature is best expressed in a series of values and principles. Anything too prescriptive would defeat the purpose of being Agile in the first place. That’s the genius of the Agile Manifesto. Scrum is a way of putting those values and principles into practice. It’s the most popular way and the best way I’ve come up with. In fact, the values are based primarily on an agreement of what Scrum and XP had in common with help from thought leaders and authors of books on software technologies.
But maybe a succinct definition of Agile could help communicate its value more clearly to the rest of the world.
I have a pretty good idea of how I would define it but I’m interested in hearing what you have to say. So here’s what I propose. Either tweet to @jeffsutherland or post below in the comments section your definition of Agile. In a twitter bound world, let’s limit it to around 100 characters (leaving some room for a # and re-tweeting) and incorporate as much of the spirit of the Agile Manifesto as necessary. It should also make sense to someone who is unfamiliar with Agile practices. The person who writes the best definition will receive three months of ScrumLab Premium for free ($150 value.)
Next week on this blog I will announce the winner and offer my definition. Tweet to @jeffsutherland and use the hashtag #IdefineAgility.
-- Jeff Sutherland
French 
English 
Spanish