BestCodingPractice Analyzer ;)

[Melba23]

boththose,

 

what is the argument for declaring outside the loop, as opposed to assigning and iterating in the same location?

I remember it being explained in another thread a long time ago along these lines:

When you declare a variable AutoIt does some checking to see if the name has been used before and in what scope. So declaring it inside a loop just causes unnecessary overhead. Although this overhead is minimal in most cases, as demonstrated by jchd's timings above, why make the CPU work harder than it needs to?

Personally I do not agree with his view that declaring the variables inside loop is "pretty useful", but as we are a broad church here and it makes so little difference, I am certainly not going to enter into an argument - there are more important things in life.

M23

[iamtheky]

 if I am going to spend more than 60ns writing the extra line i shouldve just compiled.

I agree declaring variables inside the loop is not "pretty useful", declaring them wherever i want, and in their most narrow of scopes, is the "pretty useful" part

Edited December 5, 2013 by boththose

[czardas]

For arrays created in loops, re-declaration is pretty useful and (in my view) good practice as well.

 

If the array changes size or dimension, then I recognize redeclaration as being super-efficient. If, on the other hand, an array of exactly the same dimensions is to be reused; then I see no benefit - unless you specifically require empty array elements to be used for some specific purpose. I think this is more an exception to the rule and quite unlikely to be the only solution.

Although keeping variable declarations close to their use within the code may seemingly aid readability, I think it can also add complication. Redeclaration (within a loop) hinders the flow of arguments, and. In this sense I see it as being illogical, which in turn makes the code harder to understand (to some degree ).

Edited December 6, 2013 by czardas

[jchd]

You seem to belittle the fact that a declaration in AutoIt is an initialization as well.

[czardas]

Well I'm speaking generally, and I'm always happy to reconsider my opinion. In terms of freeing up memory, I have to agree with you that it has certain benefits for larger arrays. I tend to be lazy and overwrite elements rather than reinitialize the whole array. In some instances this is indeed questionable.

Edited December 6, 2013 by czardas

[BrewManNH]

If you don't blank out the original contents of the array, how do you keep track of what's new and what's old? It's like declaring a variable in C++ but not initializing the value, you'll never know what the variable is going to hold until it's initialized with a value you set.

[jchd]

Exactly what BrewManNH said. Not all variables used inside loops are running accumulators.

What I say is that applying blindly strict directives without taking the context in consideration isn't the best practice to coding in general.

[czardas]

If you don't blank out the original contents of the array, how do you keep track of what's new and what's old?

 

It's rare that I use empty elements to keep track of anything, but I covered this scenario in post #43. Testing if an element contains data seems like a somewhat adhoc approach (although admittedly I have done this previously). I sometimes use empty array elements in other ways.

I generally keep track by keeping all the new data tight and contiguous within the array, or I try to design my code in such a way as to avoid the need to know what the elements contain. I agree this may not always be feasible.

What I say is that applying blindly strict directives without taking the context in consideration isn't the best practice to coding in general.

 

Absolutely!

Edited December 6, 2013 by czardas