|
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;
|
|
|
|
|
The first one.
Why are you even calling the second one if you're not going to use it?
-= Reelix =-
|
|
|
|
|
|
Depends on the complexity of the method containing #1. I prefer showing logic as soon as possible *in context*. I'm also a big believer in encapsulation for the soul purpose of containing the application in edible portions.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
First option, full stop.
-but-
Here's a fun little twist, if your language of choice supports this sort of thing (I'll use C# to illustrate):
void ExecuteIf(Func<bool> predicate, Action thingToExecute)
{
if (predicate())
thingToExecute();
}
ExecuteIf(() => foo, DoSomething);
...but I'm not convinced that this is any easier to read, it imposes a little extra overhead, and I would still stick with your first option. Just wanted to throw it into the mix
|
|
|
|