AbhiroopGhosh
In Mathematics (and in GMAT) the way we define squares (or perfect squares) i.e. they are number having even power of prime factors, is the definition - if I understand correctly - restricted to only integers.
You could define perfect squares (of positive integers) that way, but that would be an unconventional definition. Usually an integer is called a perfect square if it is equal to the square of an integer, i.e. if it is equal to some integer times itself. That definition has a few advantages, when you compare it with your definition; it is simpler (there's no need to know about prime factorizations to understand it) and there's no need to declare separately what is true about zero, or about one: zero and one are clearly perfect squares if we decide that a "perfect square" is the square of an integer. Then what you say, about even powers in the prime factorization of a perfect square when that square is a positive integer, becomes a consequence of the definition of "perfect square", rather than the definition itself.
This will get a bit philosophical, and is really only for interest (anyone who only cares about GMAT prep won't gain anything by reading any more of this post). In GMAT-level math, or in high school level math (which is the level of that wikipedia article you quote), when people talk about "perfect squares" they're essentially always talking about integers, and are using the definition I mention above. When you get beyond high school math, the definitions of all kinds of mathematical things are extended -- many test takers will be familiar with complex or imaginary numbers, which are defined by extending the definition of a square root so we can take roots of negative numbers, but that's just one of dozens of examples. In abstract algebra, which covers areas of math known as "group theory", "field theory", "ring theory", and more, you start to extend the meaning of mathematical operations like "multiplication" and "squaring", and the definition of "perfect square" can then extend to cover other collections of numbers, algebraic expressions, even geometric operations like rotation and reflection. You'd never even run into any abstract algebra unless you got to the second year of an undergrad math degree (or perhaps if you were learning quantum physics or some advanced computer science), so it's all way beyond GMAT scope, but in your bet, you and your friend were both right, depending on what level of math you're discussing. And if you search for "perfect square" on the internet, and look at definitions, you'll find some that support your point of view, and some that support your friend's.