Re: Hungarian notation

[pete@borland.com (Pete Becker)]

In article <4na9r2$qin@solutions.solon.com>, seebs@solutions.solon.com says...
>
>In article <3198F30F.2A2@zurich.ibm.com>,
>Keith Whittingham  <wgk@zurich.ibm.com> wrote:
>>Roger Merritt wrote:
>>> The idea is that you include information in the name of the variable
>>> to help you remember the type. So a pointer to a char might be named
>>> cpCurrent.
>
>>> Same thing applies to function names to help you remember the type
>>> returned.
>
>>And is absolutely horrible! It serves little purpose especially in the world
>>of C++ as we can invent our own data types and attempting to extend the
>>convention to include our newly defined types makes the code unreadable.
>
>(Please note that this is in several groups other than clc++.)
>
>Please don't blame Hungarian for that.  I lack the rigor and devotion
>necessary to even design, let alone use, a Hungarian style naming
>convention, but I feel compelled to defend it.
>
>It has *nothing* to do with this idiotic practice of prefixing things with
>their low-level types.  The intent was that *for any specific project* you
>would devise a naming convention which tagged things with their high-level
>type.  For instance, if you used size_t a lot, you might use the prefix s,
>so you would have sFoo for a size_t named Foo.  You would *not not not*
>encode this as ui or ul.
>
>The original Hungarian notation person has apparently written diatribes about
>M$'s systematic destruction and violation of the original idea; I'd love to
>see copies of this, so I could just mail them to people who keep quoting
>and using the utterly evil and destructive form.

Nevertheless, when someone posts a message on this newsgroup asking about 
Hungarian notation they are almost certainly referring to the utterly evil and 
destructive form, and need to be set on the right track.