I can’t tell you how many times I have had the “I want to do a complete rewrite!” conversation with developers. If you are the single, BEST coder that the world has ever seen, and I have to have the rewrite conversation with you, then you probably aren’t a good developer. I say “probably” because sometimes (rarely) a rewrite is the answer. But I am just using “rewrite” to make my point, which is to support the notion that a good developer needs common sense too. A good CODER on the other hand…

After all if a developer doesn’t know basic techniques like recursion, closures or OOP techniques it won’t be a good developer even if he’s programming in Java, PHP or C#…

That said, some languages tend to produce more effective developers even if they spend their time developing in a different language.

For example, my day job is Java (more or less). But most people we interview that only know Java are, at best, mediocre programmers, whereas those that know a more capable language tend to be both better developers, and better *Java* developers.

A good developer optimizes code. The better developer optimizes people. Code is just a means to that end. “Optimize to the max” is a sign of a developer who is applying a local optimization–”make the code as fast as possible!”, and not yet applying more global optimizations–”make the people within the business, and thereby the business itself, as successful as possible.”

Often, the slower code that is “fast enough” is better. Half a day to develop slower code vs. a full day to developer faster code means I can move on to a more important project that’s worth more money to the business. I don’t make these decisions on my own. When I started working here, I asked the business owners frequently how fast was fast enough. Now, after years working here, I’ve gotten to know how they value time vs. money and make the small decisions on my own, and only ask them the larger decisions. But the key point is that “how much to optimize” is never properly the programmer’s decision alone. To do it “just because” is spending the business’s money on things that will not return value to the business. Nothing makes a businessman happier than to realize that a programmer understands that “good enough is good enough.” The flip side of this is that when the business tells you to optimize (“this will have to be faster to handle the traffic we expect in the next six months. If you can get it twice as fast in a week, that’s money well spent.”) you’ve got a clear conscience. You’ve just been paid to play. Have fun!

As you say, you shouldn’t decide whether or not a developer is skilled by the language the developer is using. However, to pick some extreme cases, you might have good reason to question a developer who coded a web server in assembly. There is such a thing as an impedance mismatch, and even the most skilled developer might be hard pressed to overcome a bad mismatch between the tool and the job at hand. While Chuck Yeager’s saying, “It’s the man, not the machine” holds true for languages that are fairly close together in power, there does come a point where the tool’s deficiencies cannot be overcome by any amount of skill: No amount of training and skill will allow a sword yielding Samurai to defeat a man in a tank.

In this case, in my opinion, Java is not optimal for large game developments.

For example if you use PHP 5 and to choose between mysql and mysqli drivers for communicating with MySQL, the best option would be to choose mysqli, because it’s faster and object oriented code.