Talk:Orthogonal array

Question

I have problems understanding your article. For instance, you write:

Latin squares are $OA(n^{2},3,n,2)$. In order to see this, consider all triples $(i,j,s(i,j))$ where $s(i,j)$ — symbol in i-th row and j-th column in the latic [latin? PW] square. Then $n^{2}$ such triples for all $i,j\in {0,\dots ,s-1}$

but then in the example i and j run from 1 to 3 and do not start at 0. And what is s, is it 3? If so, I find it confusing that you use s for the number of "symbols" and s(i,j) for the "symbol" itself. What is the difference between n and s? Also, the term "symbol" sounds odd to me, would "object" not be a better word, or is the word symbol standard in this subfield of math?--Paul Wormer 06:42, 23 June 2008 (CDT)

The notation was a bit awkward, so I changed s(i,j) to a(i,j). And I also changed the counting so that i and j start at 0. In the Latin square example, n and s are the same. I have no idea whether "symbol" is the standard term in this field, but it may well be. -- Jitse Niesen 12:36, 24 June 2008 (CDT)

Thanks for the changes. Technically any s dictinct symbols can be used in an orthogonal array and a latin square but using ${0,\dots ,s-1}$ is more consistent. And answering previous question: n is a parameter of latin square (its size and number of symbols), s is a parameter of orthogonal array. Andrey Khalyavin 06:22, 25 June 2008 (CDT)