Coding standards
Each developer has its own view on coding standards. What about you ?
Do you prefer this syntax :
[java]int arrayIndex = 1 ;
boolean isOk = true ;
if (isOk)
{
doSomething() ;
}
else if (1 == arrayIndex)
{
doSomethingElse() ;
}
public static boolean contains (int[] anArray, int aValue)
{
for (int i = anArray.length – 1; i >= 0; i–)
{
if (anArray [i] == aValue)
{
return true ;
}
}
return false ;
}[/java]
or this one :
[java]int array_index=1;
boolean isOk;
isOk=true;
if (isOk == true) {
do_something();
} else if (array_index==1) {
do_something_else();
}
public
static
boolean
contains (int[] array, int value) {
boolean found;
for (int i=anArray.length-1; i>=0; –i) {
if (aValue==anArray [i]) {
found=true;
break;
}
return found ;
}[/java]
I’ve been using the former for a long time now (on JAVA and C/C++) and I couldn’t change my habits so easily.
A Combination of both.
public
static
boolean
contains (int[] array, int value) {
Is just plain scary. Why put all those all on a separate line?
Er.. use a tool like IDEA which can auto-format the code according to your preferences. If you don’t like the coding style, just do a re-format.
I’v seen this syntax:
public
static
boolean
contains (int[] array, int value) {
with one modifier on each line on some open-source projects. It frigthened me also! This was a reason to put this example in my post.
An IDE is not enough to enforce coding standards. If you look at the contains method, you can have multiple return or only one. A formatting tool won’t change this. It is commonly accepted that single return point is better but I sometimes find it more difficult to read.
Maybe you saw something like that multi-line signature syntax on my open-source project
.
public static
boolean
contains(
int[] array,
int value
)
{
...
}
I find this much easier to read than the one-line form. First line is modifiers, second is return type, third is name. Params each on their own line like a bullet list; no wondering about how to format a long param list.
Sorry, it is not the case that “It is commonly accepted that single return point is better”. I have blogged about it here:
http://www.jroller.com/page/matsh/20031216
I really think it is commonly accepted that single return point is better but as I said in the post, I prefer the first code with multiple return points.