Saturday, September 6, 2014

Software Physics

I am amazed at how many people claim to be better at software development than professional software project managers and engineers. Everyone knows the four dimensions of project management: schedule (time), features, resources and quality. The well understood notion is that one can fix any three dimensions while the fourth will vary. For example, management can fix the desired features, resource level and quality, however, the schedule will be whatever it will be. 

For the software professionals, we would just love our world to be this simple. The reality is that there are not only four dimensions to consider, but as many as thirty two dimensions to consider. Let me outline those that we have to balance, manage, coerce, work around, plan for, and strategize around. 

Just consider this partial list: new/emerging technologies, skill development, competitors, market evolution, technology partnership, development processes, design processes, architectures, complexity, testing/verification processes, build/code control, operational processes, collaboration, intellectual property, licensing, trouble shooting, incremental improvement, local/distributed development teaming, risk management, dependency management, economic decision making, documentation methods, customer support, requirements management, performance, scalability, reliability, talent acquisition, and customer support.

I consider each of these (and the others not listed) as fundamental aspects of software physics. A professional cannot avoid these in any project planning cycle. To avoid them, is like making a car or building that violates aspects or the laws of physics, or in other words, a disaster. I know of few professionals who willingly disregard these realities. But, why do many projects run into problems where they unintentionally violate one or more of these?

I would suggest two major reasons why software physics are ignored. The first is that executives and/or other departments believe that they know software development better than the software professions and direct them to ignore one or more aspects of software physics, maybe even unknowingly. The second is that software professionals are out of practice with mastering the management of software physics.

Let’s face it, we’re all under pressure to get more work done to ward off competitors or to increase revenues. I get why executives and other departments demand more results from software engineering. I once had a company CTO tell me that marketing would tell engineering the features, the quality level, the resource levels and schedule and that engineering’s job was to figure out how to make all of this happen. To demand than software engineering ignore one or more critical aspect of software physics means that the business is accepting one or more risks that won’t be managed, documented or visible. Wouldn’t it be better to keep the risks visible and put into place correct plans that either address or manage the risk?

While it would be nice to say that only other departments and executives force engineering to forgo acknowledging software physics, that would not be intellectually honest. I have found multiple engineering projects where aspects of software physics go unmanaged, unacknowledged and unchecked. Engineering departments are notorious for lack of transparency or, even worst, withholding reality from others.

What to do? Practice, inspect and adjust. These are fundamentals of agile development and, specifically, Scrum. Scrum’s backlog prioritization, focus on highest risk/value user-stories, and potentially shippable product completed for each sprint cycle are keys to understanding what aspects are important and those that may not apply. To master these Scrum practices means that the engineering team will also master the necessary aspects of software physics. For executives and other departments, the frequent sprint reviews provide inspection of the progress made and the opportunity to weigh in on future priorities. The frequent Scrum cycles (or sprints) means practice. 

As we know, practice makes perfect.

No comments:

Post a Comment