Best Practices

[kermitology]

A question for the developers out there.

Do you use the conditional operator in your code?

[photon]

Uh oh

[kermitology]

Hahaha.. yeah. I'm gonna kill someone.

[photon]

You're not one of those "you have to write a 1 page document justifying why you used an if statement" guys are you?

I mean I like OO and inheritance and wish I was a Smalltalk guru as much as the next guy, but still...

[kermitology]

No one ever questions the use of an if statement.

[photon]

Oh wait, or do you mean the specific ( a > b ? a : b ) thing?

[photon]

Originally Posted by kermitology

No one ever questions the use of an if statement.

Oh I've dealt with guys that want as few if statements as possible.

"if there's an if statement, that means the object model isn't correct."

[kermitology]

Originally Posted by photon

Oh wait, or do you mean the specific ( a > b ? a : b ) thing?

This.

[photon]

Ah ok, I've always called that a ternary operator.

Yeah I'm not much of a fan either. Code readability is important in my opinion and using that instead of something more explicit hurts that, though that kind of is a self defeating argument because if it was used more it would be more readable.

But I'm one who'll do:

if (a > b)
{
     doStuff();
}
else
{
     // do nothing
}

just for readability...

[amorak]

and then I ate the bowl

[BlackEleven]

Originally Posted by photon

Ah ok, I've always called that a ternary operator.

Yeah I'm not much of a fan either. Code readability is important in my opinion and using that instead of something more explicit hurts that, though that kind of is a self defeating argument because if it was used more it would be more readable.

But I'm one who'll do:

if (a > b)
{
     doStuff();
}
else
{
     // do nothing
}

just for readability...

Yes, its called a ternary operator. Some people definitely overdo it but I find it extremely useful for assignment.

Instead of

if ( y > z ) {
x = 1;
}
else {
x = 2;
}

you can just do

x = y > z ? 1 : 2;

To me, both are just as readable, but one takes 6 lines where the other takes one.

Even if you don't use it you should still know it since you'll be reading other people's code at some point and a lot of people do use it a lot, especially the old school guys.

[photon]

Yeah it's not nearly as readable to me, just cause I've never used it that much.

Plus more lines of code means my boss thinks I'm more productive!

[kermitology]

It IS a ternary operator, but that just means that it's an operator that takes three parameters.

The actual operator itself is the conditional operator.

It's far more rare to find someone who thinks that an if-else branch is bad practice than the use of that operator.

[BlackEleven]

Originally Posted by photon

Yeah it's not nearly as readable to me, just cause I've never used it that much.

Plus more lines of code means my boss thinks I'm more productive!

I see it several times a day, so I guess I'm just used to it.

I don't see why its bad practice to use it though. I actually prefer it to if/else in simple situations. It gets ridiculous for readability though when people try to do something complicated with it:

return x ? ( x > y ? 1 : 2 ) : ( x == y ? 3 : 4)

[kermitology]

Originally Posted by BlackEleven

I see it several times a day, so I guess I'm just used to it.

I don't see why its bad practice to use it though. I actually prefer it to if/else in simple situations. It gets ridiculous for readability though when people try to do something complicated with it:

return x ? ( x > y ? 1 : 2 ) : ( x == y ? 3 : 4)

OMFG.. If I EVER see a nested conditional I will seek out and destroy the person who wrote it.

It's not bad practice, but it's not good practice. I'm an extremely explicit coder. Prevents someone from doing something stupid, makes my code far more robust and clear.

[Shazam]

Originally Posted by BlackEleven

Yes, its called a ternary operator. Some people definitely overdo it but I find it extremely useful for assignment.

Instead of

if ( y > z ) {
x = 2;
}
else {
x = 1;
}

you can just do

x = y > z ? 1 : 2;

To me, both are just as readable, but one takes 6 lines where the other takes one.

Even if you don't use it you should still know it since you'll be reading other people's code at some point and a lot of people do use it a lot, especially the old school guys.

Gak. I don't like that either.

int x = 1;
if (y > z) x = 2;

I always say, if you want to stuff as much as you can into one line, use PERL. Then, after you have regained your sanity, be verbose instead.

[photon]

**** perl.

That is all.

[Mad Mel]

Originally Posted by BlackEleven

Instead of

if ( y > z ) {
x = 1;
}
else {
x = 2;
}

you can just do

x = y > z ? 1 : 2;

To me, both are just as readable, but one takes 6 lines where the other takes one.

Not all languages support it, but another way to do a very short if / else is

x = IIf(y>z, 1, 2)

A downfall in .NET is that it does evaluate both possible assignments, so both have to evaluate without error.

Personally for readability, I like Select...Case statements when they're appropriate.

Originally Posted by BlackEleven

I don't see why its bad practice to use it though. I actually prefer it to if/else in simple situations. It gets ridiculous for readability though when people try to do something complicated with it:

return x ? ( x > y ? 1 : 2 ) : ( x == y ? 3 : 4)

In SQLServer Integration Services built in objects, you have to do things this way. I frequently give myself a headache.

[Azure]

I don't know whether to cry or laugh at the fact that I'm actually getting what you guys are talking about.

[BlackEleven]

Originally Posted by kermitology

OMFG.. If I EVER see a nested conditional I will seek out and destroy the person who wrote it.

I've seen it. It's annoying, but its just readability. I'm worried about people that use memcopy, mutable variables, and non-volatile variables in threads...