greatest dilemma in the history of mankind

[wolterh]

underscores. Always.
Why? Well, as the words are separated by a '_' it is easier to determine which is the first word, and which is the second one. In that way, you can group objects by defining a parent name, for example:
login_function();
login_msg_required = "You need to be logged in to do that."

[jespdj]

It depends on the programming language I'm programming in. I use whatever is the (de facto) standard for the language. If you use something different than what almost everybody that is programming in that particular language is using, then your code will look strange to other developers.

In Java, the norm is to use camelCaseForMethodAndVariableNames and CapitalizedNamesForClasses.

In Ruby, you_normally_use_lowecase_names_with_underscores for methods and variables, and CapitalizedNamesForClasses.

[kavon89]

camelCase for methods and classes.

I tend to have underscores and all lower case variables... often I strip the vowels to make it shorter, like txt and acct.

[pp.]

On second thought, it depends on the language.

In a previous post within this thread I said I usually preferred CamelCase.

There are so-called programming languages which ignore the case in identifiers. That automatically precludes the use of CamelCase since you then could accidentially hit on a variant of a name already in use. That kind of thing would be fiendlishly difficult to debug and correct since us poor humans would find it very hard to spot this kind of conflict.

[Nemooo]

Coming from C I prefer underscores, but I generally follow the language "standard" style, so if it's Lisp im using hyphens.

[Brain-free]

Underscores for methods and variables in C.

I type fast and sometimes I forget to make a letter capital and well, the compiler responds accordingly...

[DrMega]

Coming from an old school backround, I voted camelCase.

I think underscores are better suited in constants, purely because in the olden days it was the custom to type constants in all upper case, so if they were two or more words long, it could look a bit unreadable if you didn't use underscores to break them up.

I chose camelCase as the preferred one because I think that it should be clear when scanning through the code that you are looking at a discreet item. Imagine if there is a mathematical formula and you are scanning down quickly, which of the following is more readable (in a very short time):

//camelCase

netPayThisMonth = (grossAnnualSalary - (incomeTax + nationalInsurance)) / 12;

or

//camelCase

net_pay_this_month = (gross_annual_salary - (income_tax + national_insurance)) / 12;

(incidentally, the formula is rubbish but it serves to highlight the point).

[gomputor]

CamelCase for classes and methods and underscore for the rest.
But a lot of times I just forget to do it!

[jimi_hendrix]

i use camelCase becuase my classes methods and variables tend to be one or two words

[SeanHodges]

Personally I really honestly don't mind. Picking the one that is standard in the core language API makes the decision easy and keeps things consistent, but just like the tabs vs. spaces debate I have better things to do than get religious about it...

One thing I would say though is STICK WITH YOURS OR OTHER PEOPLES ORIGINAL DECISIONS on these things. I really hate sifting through code that interleaves different coding standards


Oh, and off the record, Vim beats Emacs hands down