A lot of engineers find building applications boring, and always gravitate to working on core technologies of some kind. And certainly if pure and deep algorithms are the focus, building new core technologies is fun. But as time passes, the world of software development continues to evolve. Specifically, more and more problems have established solutions. When I think back to building educational applications in the early 1980s, we had to build our own hardware, our own operating system, our own compilers and tools, and then build the application. By the nineties this was a thing of the past, and we used off the shelf hardware, operating systems and compilers. But by now the growth of open source has changed the picture still more. Where one used to spend a lot of money on tools, now most of the tools you need are freely available and mature. And more and more of the parts you need are available too. In the early 2000s building a big Java based enterprise application required building a huge framework to do anything from Ajax to web services. Now that’s not needed either. Again, lots of open source projects to the rescue. In fact as applications become more modular, you can even get whole chunks of applications you can use. Even deployment is becoming commoditized and componentized.
So what’s left ? The answer is that the really hard stuff, what’s always been the really hard stuff is still there. You have to identify the user, understand their need and meet it. From the beginning of computing this has been the most difficult problem, because it requires us to pass from “me” to “you”. Really the mechanics of actually building an application are so much easier today than ever before, it’s crystallized the problem down to pure architecture. If you can really grasp your customers’ needs, and form a clear picture of them, modern software tools in many cases allow you to realize those needs very cleanly and transparently with surprisingly little effort. So work has decreased, but thinking is just as complex as ever. Because understanding other people has not become easier. The reason it isn’t easier is because it still requires uniting people skills, business skills and technical skills to succeed. Those skills remain hard to find in one person, both for training as well as for the underlying personality of different individuals. More than ever I feel that success in the technical world has to do with becoming a complete individual, with a well-rounded education. This requires grasping both the potentials as well as the limitations of technology, as well as the communication skills to work with customers and users, and the business acumen to understand how to do any of this without going broke.
In that sense I would encourage anyone of college age thinking about what to major in that the best thing is a combination of a good liberal arts education with in depth technical training, and some work in business and economics. It’s a combination well suited to the needs of our time, and giving people the versatility to cope with a variety of situations.
