Halbert wrote:
> How about Microsoft's engineers starting the whole Hungarian Notation 
> convention with Windows programing. 

Actually, this makes a lot of sense for C *when done right*.

The point was not to repeat in the variable name the types that you 
declared to the compiler. The point was to repeat in the variable name 
the types you CANNOT declare to the compiler.

If you have
   int line_width;
   int line_height;
   int page_height;
   int line_count;
   int pixel_count
the expressions
   page_height = line_count * line_height
and
   page_height = line_count * line_width
and
   pixel_count = page_height * line_width
and
   pixel_count = page_height * page_height
are equally type-correct as far as C is concerned, but two are nonsense.

So you declare
   int wLineWidth;
   int hLineHeight;
   int hPageHeight;
   int hwPixelCount;
   int lineCount;

Then you can write
   hPageHeight = hLineHeight * lineCount;
   hPageHeight = wLineWidth * lineCount; // Obvious nonsense!
   hwPixelCount = hLineHeight * wLineWidth; // Looks good
   hwPixelCount = hLineHeight * hPageHeight; // Nonsense!

Even picking nonsense variable names other than the notation will tell 
you that it's nonsense. If you see
   hSize = count * hQuantity; // Probably OK
   wSize = hQuantity * count; // Probably bogus

That was the idea of "Hungarian notation". Not that you should replicate 
what the compiler already knows, but distinguish types that the compiler 
can't, because you're using a language whose compiler was designed to 
fit into a PDP-11's memory.

-- 
Darren New / San Diego, CA, USA (PST)
  Helpful housekeeping hints:
   Check your feather pillows for holes
    before putting them in the washing machine.

Post a reply to this message