Seven years ago, Robin Sharp divided the lifecycle of a programming language into 7 phases:
I think Robin is correct. Once a language becomes mainstream and reaches the “acceptance” phase, it’s only a matter of time till it becomes inefficient. This is because language designers face a dilemma:
… or …
Both directions have drawbacks.
In the former case, the language doesn’t keep up with the industry. Since the language is not “state of the art”, it only adds friction and inefficiency to the craft of building software applications.
In the latter case, the language is constantly evolving to meet the state of the art, but also collecting baggage and extra weight along the way as practices fall out of fashion. Eventually the language is so large and complex it again carries an inherent inefficiency.
Since its inception, we’ve seen the addition of generics, lambdas, and dynamic programming features to C#. These are all welcome additions for me, and make the language better.
At the same time we’ve seen the rise of new paradigms for functional and multi-threaded programming. I have to wonder if keywords like lock, volatile, and delegate are out of fashion, like skinny leather ties that only clutter up a tie rack.
Do the new features in C# 4.0, like the dynamic keyword, push the language solidly into the maturation phase? Or … after 10 years … are we passed the maturation tipping point where C# can only descend towards inefficiency?
What do you think?