Agile and Beyond Shows Value of Regional Software Development Conferences
Don't overlook the value of small, regional IT events put on by nonprofit organizations: They offer candid conversations, multidisciplinary sessions and networking opportunities and, in this case, the chance to dig in to agile development.
Thu, November 15, 2012
I wasn't giving a talk, nor had my employer paid the registration fee or given me paid time off. As an independent consultants, I bill hourly, so time away from work is time unpaid. (At least it was on a Saturday—no billable hours lost.)
That said, the Agile and Beyond registration fee (about $100) covered entrance to the sessions, lunch, snacks and a chance to talk to the vendors. A half-dozen were hiring; the others were selling software. One of my clients sent a few project managers to the conference. After the event, the engineering group switched to the LeanKit Kanban project management tool based on a conversation in that vendor hall.
For that reason, and many more, Agile and Beyond—like many a small, regional, nonprofit IT conference—proved to be well worth the time, money and effort. It wasn't just me; the event sold out and attracted more than 600 attendees.
Small Conference, Big Agile Sessions
Beyond the vendors, the content included keynotes by David J. Anderson, who brought the idea of Kanban to software engineering, and Steve Denning, who wrote The Leaders Guide to Radical Management. Lanette Creamer flew in from Seattle to give a talk with Matt Barcomb, who drove up from Ohio, about the use of exploratory testing charters on an agile project.
Anderson describes agile not as an engineering concept but, rather, as a culture or "tribe." His first example was the idea of measuring the "agile-ness" of an organization in terms of the number of agile practices it uses.
Anderson's point was that such measurements have no direct connection to an actual business result or sociological outcome. To that end, he encouraged teams to consider the local context instead of adopting a cookie-cutter process.
Anderson also said that agile software development turns the old cost-of-change curve on its head. He insisted that perfect up-front-planning is expensive, often wasteful, and that it is better to make forward progress, get quick feedback and adjust than to try to prevent mistakes through perfect planning.
Meanwhile, Todd Kaufman gave a refreshing talk on agile metrics that matter with a subtitle of "How to stop abusing yourself and others with metrics." His first example was an agile team that split its work into springs and counted average velocity, or story points per iteration, to make predictions.
The problem with averages, of course, is that you are below average half the time. When the team missed its initial estimates, it needed to catch up, leading to either shortcuts that hurt future speed or, possibly, inflated estimates. Either approach hurts, rather than helps, long-term behavior. Kaufman thus recommended metrics that answer a particular question at a point in time. This is a welcome alternative to heavyweight tracking, especially if there are ways to game the system to hit the numerical target.
Finally, the exploratory testing talk exposed the elephant in the room—there are entire categories of software defects that automated tests tend to miss but are obvious to a person actually using the software. Creamer and Barcomb introduced the idea of testing charters. These are equal in weight to a development "story" and make the human tester's work visible to the entire team.
While the testing material was wonderful, the real value in the session was its multidisciplinary nature. Attendees included developers, graphic designers, managers and a handful of testers. One of the larger sources of waste I see in software is a mismatch in expectations among roles; these sorts of sessions eliminate that friction.