Skating to the puck

I recently reviewed an internal study of competing technologies for a new, large-scale, Windows application. It was the typical survey of the possible platforms for the application, summarizing the strengths and weaknesses of each. This kind of analysis has been done hundreds (or thousands, or tens of thousands) of times by companies and organizations around the world and back to the dawn of computing. The thinking is: Before we set out on this project, let's review the possible technologies and pick the best one.

As I was reading the study, I realized that the study was wrong.

Not wrong in the sense of improperly evaluating technologies, or wrong in the sense that the authors ignored a possible platform. They had included the major platforms and listed the strengths and weaknesses in an unbiased presentation.

It was wrong in the sense of time. The report had the wrong tense. It looked at the present capabilities of the platforms. It should be looking at the future.

The project is a long term project. The development is planned for five years, with a lifetime of ten years after that development. (That's a simplified version of the plan. There will be releases in the five year development phase, and enhancements and maintenance during the ten-year follow-on phase.)

For such a project, one needs a view of the future, not a view of the present. Or, as Wayne Gretzky learned from his father: "skate to where the puck is going to be, not to where it has been."

The study looked at the major technologies (.NET, Java, Silverlight, Air, and Flash) and reviewed their current capabilities. The authors made no projections of the possible futures for these platforms.

I understand that predictions are hard. (Especially predictions about the future.) But the majority of the development effort will be made in the future, from two to three years out, and continuing for a decade. Looking at the current state of technologies and deciding the next fifteen years on them is insufficient. You have to look at where each technology is going.

This study was part of a new development effort, to replace an existing product was difficult to maintain. The existing product was about ten years old, with parts going back fifteen years. The difficulties were due to the technology and design decisions that had been made at the inception of the project and during its life, some only a couple of years ago.

This team is embarking on a repetition of their previous development effort. Instead of creating a long-lasting design on robust technology, they are building a system that, in a few years, will be difficult to maintain.

Pucks move and technology changes. If you always skate to where the puck currently is, you will always be behind.