|
What if (no pun intended), the "if" actually required more complex logic, including perhaps some nested stuff, like:
if (foo != null)
{
var data = GetSomeData(foo.SomeValue);
if (data has some specific value/s)
{
DoSomething();
}
}
From an aesthetic point of view, I dislike putting all that into the main method, hence why I've got a couple "Maybe..." methods because the above scenario matches in pseudo-code what I'm actually having to deal with.
|
|
|
|
|
True there are always exceptions but for the simple example you gave I still prefer the former for the less complicated logic.
The less you need, the more you have.
Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?
JaxCoder.com
|
|
|
|
|
Marc Clifton wrote: What if (no pun intended), the "if" actually required more complex logic, including perhaps some nested stuff This is a different question. The original question is very simplistic.
If we were discussing a real scenario, I'd ask what is the conditional logic, how many times will it be repeated and in what context(s), how large is the function in which it's called, etc.
But for this simplistic scenario, I'd use #1 because it is not efficient -- either in coding and probably execution -- to call a function for no useful reason.
|
|
|
|
|
If the test is complex and gets used more than once, there's an argument for giving the test its own method. That way the test is isolated, and can be changed later, but it's not hidden away in the MaybeDoSomething method. So you might end up with
if (TestIfWeNeedToDoSomething( ... ))
{
DoSomething();
}
// lots of code follows.
if (TestIfWeNeedToDoSomething( ... )) // again
{
DoSomething();
}
and so on. A little inelegant, though.
|
|
|
|
|
That is a different situation than posed. The overhead of a function call could be justified in this case.
|
|
|
|
|
I'm with all the others: first version is what I'd use.
It's easier to read, more efficient, and potentially means you don't have to carry the "decision variable(s)" through to the called method.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
the first. Almost without exception always.
Even if the logic starts to get complex. It will still be understandable to read. PAinful potentially. but the second is harder to read and the more complex it gets the better in comparison to the first, but the first will always be more readable.
IMO. ymmv. and all that quid pro quo that goes with all that.
To err is human to really elephant it up you need a computer
|
|
|
|
|
|
Both have a usage. If it's code that is going to be reused somewhere else, version 2. If it's lots of crazy complicated code, version 2. If it's pretty simple and not reused, version 1.
|
|
|
|
|
|
Agreed!
|
|
|
|
|
To me, it mainly depends on whether foo is incidental to (option #1), sufficient for (option #2), or necessary for (option #3) the execution of DoSomething() . Option #3 would be refactoring DoSomething() to include foo so not always an option.
So if foo gatekeeps the execution of DoSomething() in this one spot but not others then it's incidental. If foo gatekeeps DoSomething() in a lot of spots but not all then it's sufficient. If foo always gatekeeps DoSomething() then it's necessary.
|
|
|
|
|
I much prefer the first method.
Having said that I have been trying to dis the second method and the only things I can say is it looks ugly and is doing more than one thing (both testing if something can be done and doing it.)
It has the advantage that the test is explicit that it should be done.
And yes, I have well over-thought this...
-showing my working out-
Renaming the foo variable to suit what it is being used for:
if (canDoSomething)
{
DoSomething();
}
or:
DoSomethingIfPossible(canDoSomething);
...
DoSomethingIfPossible(bool canDoSomething)
{
if (canDoSomething)
{
}
}
Or, if renaming the variables is not viable you could try these three convoluted options:
if (CanDoSomething(foo))
{
DoSomething();
}
bool CanDoSomething(bool canDoSomething)
{
return canDoSomething;
}
or:
DoSomethingIfPossible(CanDoSomething(foo));
...
DoSomethingIfPossible(bool canDoSomething)
{
if (canDoSomething)
{
}
}
bool CanDoSomething(bool canDoSomething)
{
return canDoSomething;
}
or:
DoSomethingIfPossible(foo);
...
DoSomethingIfPossible(bool foo)
{
if (CanDoSomething(foo))
{
}
}
bool CanDoSomething(bool canDoSomething)
{
return canDoSomething;
}
|
|
|
|
|
I like the "IfPossible" version as well. I'll have to remember that.
|
|
|
|
|
How about:
#define DOSOMETHING_IF(A) if((A)) DoSomething()
- No extra calls if condition is false.
- No repeated if statements visble
Mircea
|
|
|
|
|
There are only three acceptable reasons for creating a function:
1. If there is more than one occurrence of the segment of code.
2. It makes reading the code more logical.
if(itShouldBeDone(foo)) {
doSomething()
}
3. to confuse the next poor soul who gets to work on your code.
function itShouldBeDone(foo){
return decideWhatToDo(foo);
}
function decideWhatToDo(foo){
let decision= false;
if(typeof foo == "boolean") {
switch(foo){
case true:
decision= foo;
break;
case false:
decision= !foo;
break;
default:
decision= true;
}
}
return false;
}
Nothing succeeds like a budgie without teeth.
|
|
|
|
|
Definitely the first one.
Low level functions should only do things, preferably in a class that can get mocked during testing.
High level functions should decide which things are worth doing, and are prime targets to test.
|
|
|
|
|
The first one.
"In testa che avete, Signor di Ceprano?"
-- Rigoletto
|
|
|
|
|
I prefer the first as the DoSomething() function is explicit what it does.
just like the second it encapsulates all actions to be done, that is what functions are made for.
The second MaybeDoSomething(bool foo) function is less explicit.
Is the "maybe" related to the bool parameter? Suppose so, but that is an assumption, or?
And will it "doSomething" if the parameter is false or when the bool is true?
For me the first construct does not have these question and is therefor more clear.
|
|
|
|
|
It depends. If The Something is part of the business logic I would encapsulate it, if it is simple flow control I wouldn't.
Encapsulating in a different subroutine offers better options for bugfixing, bug reporting, logging, maintenance, refacotring, rewriting, whatever.
GCS d--(d-) s-/++ a C++++ U+++ P- L+@ E-- W++ N+ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
|
|
|
|
|
Option 1 unless there is a need for re-use. Going back to my days in machine code when every clock cycle counts then the overhead of the call to a method would leave me no choice but to use option 1.
|
|
|
|
|
Depends on context but esp. in complex scenarios I like the second form because it allows for having clearer understand of the flow while being composable:
I could have a bigger methods which calculates certains states and then calls several MayBe-methods...
It allows the MayBe-method to be called based on the state which in turn can be easily serialized and deserialized (which is not so easy with an "if"-Statement).
|
|
|
|
|
from a debugging point of view the second is a nightmare, you need to step into the maybe to know if it happend or not. I will take the first code any time
|
|
|
|
|
this reminds me of a code review comment I received, for using:
while (true)
{
failure_code = 123;
if (failure_condition1) break;
failure_code = 125;
if (failure_condition2) break;
...
failure_code = 0; break;
}
if (failure_code>0)
process_error(failure_code);
else
process_policy;
To me, the fake structure of the while loop supported the early exit (once we know we failed)
The code was so overly complicated (insurance rules), that I could only prove correctness by literally applying the "rules & regs" line by line!
For what it's worth, the Business Analyst could understand, and the programmers complained that the while loop was throwing them off. [Comments were added]
|
|
|
|
|
I don't like the idea of having a DoSomething() and a MaybeDoSomething that only adds a simple if {...} around a call. My approach would have been:
void DoSomething(bool something_allowed)
{
if (something_allowed)
{
}
} I've worked with code that was written such that there is only one or two block constructs per method. Naming ends up stupid (OpenFileIfFoundAndNotAlreadyOpen() , anyone?), and it makes it very difficult to follow logic. I would much rather have a method run a little longer yet be able to see the entire algorithm in one place.
Software Zen: delete this;
|
|
|
|