#1. iterator and book are different types. You can do ++iterator but with book… Does it make sense? So a book is not an iterator and iterator is not a book it is another type!

#2. Yes sometimes it is much more expressive to give a variable reasonable name, but sometimes temporary is just temporary storage (i.e. backup means reserved copy and tmp is not a reserve it is just temporary storage)

#3. And again: it depends 🙂 function name nthPowerOf2(int n) is a little bit more expressive then nthPowerOfTwo(int n) (IMO), but otherwise i agree, that it is bad practice in general.

#5. I thought it is well-established term (like file handle, window handle etc). Wiki: In computer programming, a handle is an abstract reference to a resource.

#6. As Miguel has already mentioned, one-letter variables are great in math code or array traversing (again, it is well-established practice)

#7. For example: std::atomic is pretty much self-explanatory and simple, while std::atomic is too redundant. But i agree with you when dealing with complex templates (i.e. more than one type, or type with some restrictions).

By the way, thanks for a great articles! Keep up the good work! 😉

For example, struct Point2d { double x, y; } is a pretty good naming convention for, well, the x coordinate and the y coordinate. I’d hate using “abcissa” and “ordinate” or some other weird name like that.

Another good place to use single letter names is that i and j are usually the indices for traversing an array. And don’t say “use range-based for or the STL”. Sometimes you need the indices and sometimes you need to manipulate the indices.

Anyway, the other ones are fine. Nice video.