7 Programming Myths: Busted!
Unfortunately, the blame for this myth falls on Fred Brooks himself. Well, almost -- he's been misquoted. What Brooks actually says is that, in one study, the very best programmers were 10 times more productive than the very worst programmers, not the average ones.
Most developers fall somewhere in the middle. If you really see a 10-fold productivity differential in your own staff, chances are you've made some very poor hiring choices in the past (along with some very good ones).
What's more, the study Brooks cites was from 1966. Modern software project managers know better than to place too much faith in developer productivity metrics, which are seldom reliable. For one thing, code output doesn't tell the whole story. Brooks himself admits that even the best programmers spend only about 50 percent of the workweek actually coding and debugging.
This doesn't mean you shouldn't try to hire the best developers you can. But waiting for superhuman coders to come along is a lousy staffing strategy. Instead of obsessing over 10x developers, focus on building 10x teams. You'll have a much larger talent pool to choose from, which means you'll fill your vacancies and your project will ship much sooner.
Programming myth No. 4: Cutting-edge tools produce better results
Software is a technology business, so it's tempting to believe technology can solve all of its problems. Wouldn't it be nice if a new programming language, framework, or development environment could slash costs, reduce time to market, and improve code quality, all at once? Don't hold your breath.
Plenty of companies have tried using unorthodox languages to outflank their competitors. Yammer, a social network, wrote its first version in Scala. Twitter began life as a Ruby on Rails application. Reddit and Yahoo Store were both built with Lisp.
Unfortunately, most such experiments are short-lived. Yammer switched to Java when Scala couldn't meet its needs. Twitter switched from Ruby to Scala before also settling on Java. Reddit rewrote its code in Python. Yahoo Store migrated to C++ and Perl.
This isn't to say your choice of tools is irrelevant. Particularly in server environments, where scalability is as important as raw performance, platforms matter. But it's telling that the aforementioned companies all switched from trendy languages to more mainstream ones.
Fred Brooks foresaw this decades ago. In his essay "No Silver Bullet," he writes, "There is no single development, in either technology or management technique, that promises even one order of magnitude improvement in productivity, in reliability, in simplicity."
For example, when the U.S. Department of Defense developed the Ada language in the 1970s, its goal was to revolutionize programming -- no such luck. "[Ada] is, after all, just another high-level language," Brooks wrote in 1986. Today it's a niche tool at best.
Of course, this won't stop anyone from inventing new programming languages, and that's fine. Just don't fool yourself. When building quality software is your goal, agility, flexibility, ingenuity, and skill trump technology every time. But choosing mature tools doesn't hurt.
Programming myth No. 5: The more eyes on the code, the fewer bugs
Open source developers have a maxim: "Given enough eyeballs, all bugs are shallow." It's sometimes called Linus' Law, but it was really coined by Eric S. Raymond, one of the founding thinkers of the open source movement.
"Eyeballs" refers to developers looking at source code. "Shallow" means the bugs are easy to spot and fix. The idea is that open source has a natural advantage over proprietary software because anyone can review the code, find defects, and correct them if need be.
Unfortunately, that's wishful thinking. Just because bugs can be found doesn't mean they will be. Most open source projects today have far more users than contributors. Many users aren't reviewing the source code at all, which means the number of eyeballs for most projects is exaggerated.
More importantly, finding bugs isn't the same as fixing them. Anyone can find bugs; fixing them is another matter. Even if we assume that every pair of eyeballs that spots a bug is capable of fixing it, we end up with yet another variation on Brooks' Mythical Man-Month problem.
One 2009 study found that code files that had been patched by many separate developers contained more bugs than those patched by small, concentrated teams. By studying these "unfocused contributions," the researchers inferred an opposing principle to Linus' Law: "Too many cooks spoil the broth."
Brooks was well aware of this phenomenon. "The fundamental problem with program maintenance," he wrote, "is that fixing a defect has a substantial (20 to 50 percent) chance of introducing another." Running regression tests to spot these new defects can become a significant constraint on the entire development process -- and the more unfocused fixes, the worse it gets. It's enough to make you bug-eyed.
Programming myth No. 6: Great programmers write the fastest code
A professional racing team's job is to get its car to the finish line before all the others. The machine itself is important, but it's the hard, painstaking work of the driver and the mechanics that makes all the difference. You might think that's true of computer code, too. Unfortunately, hand-optimization isn't always the best way to get the most performance out of your algorithms. In fact, today it seldom is.
One problem is that programmers' assumptions about how their own code actually works are often wrong. High-level languages shield programmers from the underlying hardware by design. As a result, coders may try to optimize in ways that are useless or even harmful.
Take the XOR swap algorithm, which uses bitwise operations to swap the values of two variables. Once, it was an efficient hack. But modern CPUs boost performance by executing multiple instructions in parallel, using pipelines. That doesn't work with XOR swap. If you tried to optimize your code using XOR swap today, it would actually run slower because newer CPUs favor other techniques.
Multicore CPUs complicate matters further. To take advantage of them, you need to write multithreaded code. Unfortunately, parallel processing is hard to do right. Optimizations that speed up one thread can inadvertently throttle the others. The more threads, the harder the program is to optimize. Even then, just because a routine can be optimized doesn't mean it should be. Most programs spend 90 percent of their running time in just 10 percent of their code.
In many cases, you're better off simply trusting your tools. Already in 1975, Fred Brooks observed that some compilers produced output that handwritten code couldn't beat. That's even truer today, so don't waste time on unneeded hand-optimizations. In your race to improve the efficiency of your code, remember that developer efficiency is often just as important.
Programming myth No. 7: Good code is "simple" or "elegant"
Like most engineers, programmers like to talk about finding "elegant" or "simple" solutions to problems. The trouble is, this turns out to be a poor way to judge software code.
For one thing, what do these terms really mean? Is a simple solution the same as an elegant one? Is an elegant solution one that is computationally efficient, or is it one that uses the fewest lines of code?
Spend too long searching for either, and you risk ending up with that bane of good programming: the clever solution. It's so clever that the other members of the team have to sit and puzzle over it like a crossword before they understand how it works. Even then, they dare not touch it, ever, for fear it might fly apart.
In many cases, the solution is too clever even for its own good. In their 1974 book, "The Elements of Programming Style," Brian Kernighan and P.J. Plauger wrote, "Everyone knows that debugging is twice as hard as writing a program in the first place. So if you're as clever as you can be when you write it, how will you ever debug it?" For that matter, how will anyone else?
In a sense, concentrating on finding the most "elegant" solution to a programming problem is another kind of premature optimization. Solving the problem should be the primary goal.
So be wary of programmers who seem more interested in feathering their own caps than in writing code that's easy to read, maintain, and debug. Good code might not be that simple. Good code might not be that elegant. The best code works, works well, and is bug-free. Why ask for more?