Have any identifier case readability studies been done?

Question

Different languages have different standards for identifier casing. Most discussion I see about this subject revolves around a specific language, and is loaded with personal opinions. And these strong opinions all stem from the decision of one person or a group of people who created the language.

I'd love to see a graph showing just how much more readable words_per_minute is than wordsperminute. Obviously I'm most interested in the more controversial cases such as wordsPerMinute. I suspect the most readable ones take the most time to type, and vice versa, but of course I can't be sure of this on my own.

Have there been any language-agnostic studies done comparing readability of different capitalization schemes? Stats on speed of typing would also be welcome.

Answer 1

Your question reminds me of this blog post. Title: Readability, an experimental view

Answer 2

I've seen a few citations, but tracking down real studies seems to be quite difficult. I've seen only one or two available online that seem to have any real supporting evidence.

Answer 3

I recall there was a paper on that topic at ICPC 2009 in Vancouver, but I was not convinced by their evaluation approach.

The PDF can be found here: http://www.cs.loyola.edu/~binkley/papers/icpc09-clouds.pdf

They showed a screen with moving words and folks at to click as fast as possible on given sentences (once using underscores and once camel-case). This is (sorry to say) a pointless evaluation approach for several reasons, among them that source code does not move, and that it would be much more interesting to learn about the readability in the context of source code with all its operators and other special characters!

PS: interestingly enough there are some languages that use case as part of their syntax! Haskell, Ruby, and Self are among those I know of.