Almost five and a half years ago, one my first project in the UK, I built a constrained email tempting engine to be used by marketing folk with a small team of three other developers. I think we worked for about six to eight weeks to develop the entire application, including retrofitting its integrate its output with another system.

A year or two ago, we had some consultants return to the client who talked to one of the original developers who still worked for the same client. They spoke fondly of the good times he had on our team, and spoke proudly about the application we wrote. In the four years or so of constant use (they would revisit their emails at least once a month), the users apparently only ever reported one bug. He mentioned that bug (a special character in the email subject line) was also deprioritised from the original work we did. It was an enough fix and with test around it, was not a problem updating the system.

I certainly appreciated the feedback, particularly since you don’t always get to return to see the long term results of the work you do as a consultant, or even just as a developer.

Some of the great things I learned from that project.

Acceptance Test Driven Development is entirely possible
We worked hard to test drive this Swing based application from the start. It really changed the way that we designed the application and all of us learned a whole heap about understanding testable presentation patterns. It formed the basis of several talks and the knowledge, particularly the separation of concerns is entirely applicable to all sorts of other tools and interfaces such as javascript, or command line interfaces.

On a different note, coverage combining end to end acceptance test and unit tests meant we had 100% code coverage. It was never the goal, but interesting to see on this application it was entirely possible.

Never discourage people from trying new things
The organisation was rife with contractors. Not all of them entirely passionate. I love building great teams and the fact that everyone was learning, and saw things being delivered reignited people’s passions to do various things. I distinctly remember two events made possible by encouraging people to explore their passions and strengths.

One of the developers, interested in the User Experience side asked if he could do some user testing with the application. I can’t remember exactly what I said, but certainly encouraged him to do that. He did, what some people call Guerilla Usability Testing – grabbing someone who didn’t know anything about our application, and giving them a task list to complete as he watched them over the shoulder. Based on this study, we found out, to no surprise, the way we laid out the application didn’t make some of the tasks as easy as they could have been. We channeled this feedback into our interface and helped make life easier for its users.

The other event, a contractor, seemed to find his passion for developing usable software reignited. Everyone will know that swing applications aren’t the nicest things to look at. I remember coming in one morning to find that whole interface reskinned and redeveloped. It looked a whole load better than what you’d get with swing out of the box.

I found out much later that he had, by his own volition, worked a couple of extra hours one night to reskin our interface using JGoodies because he wanted make the interface attractive to users. Awesome stuff that would never have happened without encouragement to do “the right thing”.

Involving real users
As a systems thinker, I know that the purpose of the project should have been an intermediate step towards a bigger change. The real problem was that the organisation was a missing feedback loop from IT back into marketing. I knew that then, and I know that know, however being pragmatic and realistic, you can only do good things within your own sphere of influence. Trying to change how the entire marketing and IT department silos worked (or failed to work together) would take much more than the two or three months I was going to be there.

Fortunately, we had one person on our project who had worked for their company for a while and had a couple of good contacts in marketing. Whilst we couldn’t change the way projects spun up, we managed to get someone who was going to use our software down, early, and show them what we were building for them. More than that, they asked for somethings (that we could almost immediately turn around for them) and they were blown away that their feedback actually mattered and made a difference.

I appreciate good projects, and where possible, create those environments for others to thrive in. There are many good things to carry forward in these things and can only hope others benefit from it as well.