Founding Hungarian notation Documentation

Just because the inventor is an amateur astronaut doesn't mean it's worth using. just say no to Hungarian notation! Unless (perhaps) you're an assembly language programmer that needs some way of documenting what types your untyped registers hold.

Here's the reasons:

- your code becomes far less readable by reducing abstraction - types suddenly become how the data is stored and not what their capabilities are

- there is no standard for Hungarian notation - your lpszName is someone elses strName and someone elses c_strName.

- it lies - when you see lp<anything> it doesn't mean a selector/offset pair

- even the company that created it don't like it anymore - .Net APIs seem to get away without needing types encoded in variable names

Cheers,

Ash

Hungarian notation had is motivation when all programming was mainly made in C, due to its weak type-system. It was meant to let you know the type of a variable to avoid you to invoke for it operations that where meant for other types of variables. In these cases the compiler says nothing about, and you will discover the bug may be after selling your app to 1000000 users!

If you use C++ properly (not as a "C with steroids", but with proper function overloading, templates and generic algorithm) that kind of things cannot happen (simply, the code doesn't compile thus revealing immediately the mistake) and also, "types" may be misleading, since a same generic algorithm may work with different types (thus you go with laying names...)

Of course, it still have sense when C++ code invokes plain old C APIs, but such part of code should be localized in well defined modules and not distributed everywhere.
Good design doesn't need to be Hungarian.