Which do you guys prefer to use as far as Win32 programming and Win32 GUI development?  MFC or traditional Win32 API calls, etc.  I'm asking because I'm just trying to get an idea of what people typically use, and perhaps why people would use one method over the other and which advantages one brings over the other if any, and lastly when should you decide to use one method over the other in various situations.

MFC is rather VB-like: everything is an object, and quite a bit of really low level window management is automated. If you like that sort of environment, it's not that bad. It does entail some extra overhead (both in terms of work required and program size) but it isn't nearly as bad as VB.

I generally prefer to do it myself, but I will use MFC if I have to.

I've noticed that, at least from the MFC specs, MFC is mostly API calls minus one or two parameters that are part of the function call.  For example, SetWindowText on a CWnd object simply calls the SetWindowText API method, passing along the member handle variable.

I like Win32.