Mmm.. I didn't read your original words clearly with regard to the component internals; seems we agree. It's a point worthy of being re-iterated numerous times though; that it's more important to maintain consistency than conform to someone else's practices.

With type prefixing; the developer at some point has to make a reasoned abstraction of the variable's type description - and that abstraction has to facilitate flexibility; there's no point calling a variable iNum if it might be changed to a long-integer at definition later; and prefixing all long's as 'i' to avoid that problem is obscuring type information informally, that would pertain to lesser readability, even confusion.

That was a somewhat contrived example. In a perfect system; there wouldn't be many names of fully qualified classes sitting around in hard-coded routines; but unplanned systems often start out imperfect =P.