Lanning Blog

Hard Earned Statistics

On comment density

Capers Jones points out that studies at IBM found that a commenting density of one comment roughly every 10 statements was the density at which clarity seemed to peak. Fewer comments made the code hard to understand. More comments also reduced code understandability (Jones 2000).
This kind of research can be abused, and projects sometimes adopt a standard such as “programs must have one comment at least every five lines.” This standard addresses the symptom of programmers’ not writing clear code, but it doesn’t address the cause.
(Code Complete 2nd Edition)

On module size

In nonmathematical terms, Hatton's empirical results imply a sweet spot between 200 and 400 logical lines of code that minimizes probable defect density, all other factors (such as programmer skill) being equal. This size is independent of the language being used — an observation which strongly reinforces the advice given elsewhere in this book to program with the most powerful languages and tools you can. Beware of taking these numbers too literally however. Methods for counting lines of code vary considerably according to what the analyst considers a logical line, and other biases (such as whether comments are stripped). Hatton himself suggests as a rule of thumb a 2x conversion between logical and physical lines, suggesting an optimal range of 400–800 physical lines.
(The Art of Linux Programming)

It appears to be possible to build significant systems (FitNesse is close to 50,000 lines) out of files that are typically 200 lines long, with an upper limit of 500. Although this should not be a hard and fast rule, it should be considered very desirable. Small files are usually easier to understand than large files are.
(Clean Code)

On variable name length

Gorla, Benander, and Benander found that the effort required to debug a program was minimized when variables had names that averaged 10 to 16 characters (1990). Programs with names averaging 8 to 20 characters were almost as easy to debug. The guideline doesn’t mean that you should try to make all of your variable names 9 to 15 or 10 to 16 characters long. It does mean that if you look over your code and see many names that are shorter, you should check to be sure that the names are as clear as they need to be.
(Code Complete 2nd Edition)

On programming language choice

Comparison of Markov Chain implementations
Language, Lines of source code
C, 150
Java, 105
C++/STL/list, 70
AWK, 20
Perl, 18

(The Practice of Programming)

blog platform doesn't support standard markdown tables, sorry.

Programmers are more productive using a familiar language than an unfamiliar one. Data from the Cocomo II estimation model shows that programmers working in a language they’ve used for three years or more are about 30 percent more productive than programmers with equivalent experience who are new to a language (Boehm et al. 2000). An earlier study at IBM found that programmers who had extensive experience with a programming language were more than three times as productive as those with minimal experience (Walston and Felix 1977). (Cocomo II is more careful to isolate effects of individual factors, which accounts for the different results of the two studies.) Programmers working with high-level languages achieve better productivity and quality than those working with lower-level languages. Languages such as C++, Java, Smalltalk, and Visual Basic have been credited with improving productivity, reliability, simplicity, and comprehensibility by factors of 5 to 15 over low-level languages such as assembly and C (Brooks 1987, Jones 1998, Boehm 2000).
(Code Complete 2nd Edition)

On bugs

The fundamental problem with program maintenance is that fixing a defect has a substantial (20-50 percent) chance of introducing another. So the whole process is two steps forward and one step back.
(The Mythical Man Month)

Industry average experience is about 1–25 errors per 1000 lines of code for delivered software.
(Code Complete 2nd Edition)

On chunks of information

Miller (1956) summarized evidence that people can remember about seven chunks in short-term memory (STM) tasks. However, that number was meant more as a rough estimate and a rhetorical device than as a real capacity limit. A single, central capacity limit averaging about four chunks is implicated along with other, noncapacity-limited sources.
(Cowan 2001)

On choices

Hick's law, or the Hick–Hyman law, named after British and American psychologists William Edmund Hick and Ray Hyman, describes the time it takes for a person to make a decision as a result of the possible choices: increasing the number of choices will increase the decision time logarithmically. The Hick–Hyman law assesses cognitive information capacity in choice reaction experiments. The amount of time taken to process a certain amount of bits in the Hick–Hyman law is known as the rate of gain of information.

- 4 toasts