Thread: is prefixing by Hungarian notation important?

Hi!

Right now I'm trying to teach myself a standard way of coding. That includes how to annotate function and variable names, where to put what in a file etc.

The other day I found this document. What struck me was this:

Variable names should be prefixed by Hungarian notation.

Do I really need to do that? If you look at the table of the prefixes listed on that page, all my class vars should like:

Code:

int m_iSomeVar;
float m_fSomeFloat;

I have never prefixed my vars before, so to me it looked a little odd.

My question is: Do I need to using this kind of prefixing, or is it OK to continue without it?

Thanx alot

Well a board search will show up what I think about it.

No, it's not important IMO.

It should be avoided like the plague.

Why is hungarian notation so bad? It let's you know what type variables are without having to go back and look them up.

How does the programming industry view hungarian notation? Does your workplace use it?

http://cboard.cprogramming.com/showt...light=notation

> Why is hungarian notation so bad?
1. It destroys readability by putting a bunch of garbage on the front of every symbol.
2. You rapidly run out of meaningful single letters - does 'u' mean unsigned or union for example?

> It let's you know what type variables are without having to go back and look them up.
Sure, if everyone uses the rules properly, but how often does that happen. Moreover, there are many local variations, so even if you manage to learn one set of rules, you have no information when you look at someone elses code from a different environment.

Besides, I say if your function no longer fits on screen, you have way more problems than simply not knowing what the types are.

> Does your workplace use it?
The fact that my previous place used it so badly was a contributing factor to me leaving for somewhere more sane.

Originally Posted by Salem

Originally Posted by The Brain

It let's you know what type variables are without having to go back and look them up.

Sure, if everyone uses the rules properly, but how often does that happen. Moreover, there are many local variations, so even if you manage to learn one set of rules, you have no information when you look at someone elses code from a different environment.

I agree with Salem. Since C++ is a strongly-typed language anyway, there's no real reason to keep type information in the variable name. It also makes it a hastle if you ever have to change the type of one of your variables.

However, I like Hungarian notation. It can be very useful if used correctly. Take a look at the this article.

For what it's worth, I prefer to use descriptive names rather than hungarian notation to try to keep track of what a variable is and it's type. I also find significant use of typedefs, ala Windows, to be very distracting. I'm also at odds opinions such as:

>> Besides, I say if your function no longer fits on screen, you have way more problems than simply not knowing what the types are.

I personally find that reading through a long function is easier than jumping back and forth between a multitude of functions, particularly if the goal is to keep the functions short, rather than reuse.

I certainly understand the rationale for use of hungarian notation, typedefs, and reasonable size functions. I believe that each programmer and each employer decides for themselves where their comfort zone is.

If you ever need to know the type of something just hover over it in the IDE and the IDE will let you know. There is never any need for type information in variable names.

yep read it couple times, still dont agree. With modern IDE's there just isnt any need whatsoever to obfuscate your own code unnecessarily and thats exactly what hungarian notation does, it destroys any readability the code has and for that reason alone it should not be used. It really doesnt do anything useful for you at all.

Originally Posted by Dante Shamest

. Your IDE might tell you it's an int, but it won't tell you what kind of thing it's measuring (length of something? a counter in a loop?).

Code:

int iHead; // index in something
int cShots; // a counter

I fail to see how anyone would assume cShots meant "Counter for Shots", and not "A char variable named cShots" without looking at the word "int", which automaticly makes the use of hungarian notation pointless right there.

Descriptive names are for more better:

Code:

int HeadIndex; //index in something
int ShotCounter; // a counter

It isn't that much more to type, and far more descriptive.

heard of typedef??