Blogging Roller: The thing that killed hungarian.

The thing that killed hungarian.

Shawn A. Van Ness quoted by Brad Wilson: The biggest single thing that killed hungarian within MS, for .NET, is the lack of consistency with which it's been used. That's the one thing that both the proponents *and* opponents of hungarian could agree on: the single worst variable naming convention in the world is "inconsistent hungarian". And so, nobody could come up with a perfectly consistent (yet reasonably consice) system for applying hungarian notation to such a richly typed world as .NET (viz. what's the prefix for ApplicationException?). So it's gone. End of story.

RIP.

Dave Johnson in Microsoft • 🕒 07:01AM Jan 08, 2003

Tags: Microsoft

Comments:

That didn't kill hungarian. That killed the USE of hungarian, because people always seem to assume that hungarian applies across problem domains, when it doesn't. Hungarian was meant to apply prefixes BY PROBLEM DOMAIN - so cFoo meant "colour [of] foo" if it was a colour, or "character representing foo" if it was a wide character... people always assumed the prefixes meant raw data types, which is wrong. Worse, most books took the short path and said that the prefixes WERE data types, making it worse on ol' Hungarian. (That said, I hate Hungarian, because I use strongly-typed languages where it's unnecessary.)

Posted by Joseph Ottinger on January 08, 2003 at 03:36 PM EST #