Help me out with this code please!

Quote:

---

Originally posted by sleeping
It is my opinion that (almost) everybody uses single-character identifiers for for-loops, and the widely spread identifiers i, j and k for loops shouldn't confuse anyone.

---

If it is a simple loop, then I don't think that short names bother people. It's when loops get to be long and complicated.

However, take a look at the following code from Numerical Recipes in C online PDFs:

amotsa() on page 11 of Chapter 10 Section 9

pages 9 and 10 of Chapter 9 Section 7

I have used seen rather complicated loops in code I have worked with -- on the orders of 500+ lines with 4 embedded loops. In here, clear variable names are critical to the understanding of the code.

- Kevin

Before anyone says "they should have broken up that code into multiple functions", I want to state that it was C code (no inlining) and was necessary to optimize the speed of the code (it was very expensive to push and pop things off the stack and switch contexts).

Caution - religious wars! Why is it that the most banal, arcane subjects always get discussed in the most controversial manner?

Just a few things to clarify my last post: Nobody said that short identifiers are bad - it all depends on the problem domain, not on whether they are used as loop counters or not.

Just consider the following examples:

Code:

---

// Call DoSomething() 42 times:
for(int i = 0; i < 42; i++)
{
  DoSomething();
}

---

In this example, 'i' has no other meaning than being a simple loop counter. A valid name might be 'count', but just as well 'n' or 'i'. It doesn't really matter, so using a generic name as 'i' is fine, especially as the loop counter isn't used for anything else inside the loop.

Now consider the following example, which might be part of an algorithm for some kind of matrix transformation:

Code:

---

for(int i = 0; i < iMax; i++)
{
  for(int j = 0; j < jMax; j++)
  {
    matrix[ i][j] = SomeCalc(i, j);
  }
}

---

In this case, too, names like 'i' and 'j' are fine, as they are commonly used as matrix indices in mathematics. A mathematical paper describing that transformation is likely to use 'i' and 'j' as well, so there's a direct mapping between the terminology used in the problem domain and the code. Likewise, variable names like m, n; p, q, r; x, y, z are fine for code which directly maps to common mathematical notation.

But now consider the following example, where something has to be done for every month in a given range of years:

Code:

---

for(int year = firstYear; int year <= lastYear; year++)
{
  for(TMonth month = january; month <= december; month++)
  {
    // Do something with 'year' and 'month'
  }
}

---

In this example, it would be completely silly to use 'i' and 'j' instead of 'year' and 'month', since the loop variables clearly stand for years and months. Using something like 'i' and 'j' in the code inside the loop would obfuscate the real meaning of those variables. And it's the same in the OP's code: Since the loop counters and array indices have the meaning of 'StudentId', 'CategoryId' and 'GradeId', the most natural thing is to use those names instead of 'i', 'ii' and 'iii'.

The bottom line: Just because there are many examples where one-character names like 'i' and 'j' make sense (usually in the role of variable names in a mathematical domain or as generic indices without further meaning), that doesn't imply that counters in a for-loop must always be that short - it all depends on what they actually represent.