It just wouldn't work, either. A number is a very static thing. You can point to it and say: "It is exactly x divided by y plus z".. or whatever. Infinity is very different from that.. It isn't the biggest number in existence... You will never reach infinity if you continue adding numbers to eachother. It isn't a number, you couldn't use it in equations as one.
I disagree completely. A number is a vague thing, difficult to define. There are different things that may be called infinite, and using infinity as a number may mean that it cannot be used to refer to all kinds of things that may be called infinate. But it can be used to measure, indicate quantity, or convey a particular ratio, all which are things that are ascribed to numbers. It can be a number.
You can't reach i if you continue adding integers to each other either, but we consider i a number. It's not a "Real" number, and unlike any Real number it does not have express an ordering. Infinity is similar. It's not "Real" but unlike i it expresses an ordering. Unlike 1.3 it can express the number of objects in a group.
A number must generally be considered together with the set of numbers it is a part of. So 1 is an integer, but it is also a complex number, and when you think of one as a complex number, you cannot use it to express relational order. Likewise we may think of the set of integers modulo 256 as a set in which "1" has a special meaning and is a number. And just as we can define integers modulo 256, which happens to be how the smallest unsigned numbers in a computer are conceived, we can define IEEE 754 double precision floating point numbers. IEEE 754 double precision floating point numbers can express values in the non uniform range [2.226*10^-308,17976*10^308], plus 0, -0, Inf, -Inf, and NaN. Why should these not be called numbers? They are surely the most used kind of number in modern scientific computation. IEEE 754 defines one set of properties for infinity, with particular rules intended to make them useful to computer programmers and users.
Another set of numbers for which infinity exists are Hyperreal numbers. Hyperreals help realize the distinction between the sum 10^i, and the sum the sum 2^i. The latter sum converges toward infinity much more quickly, as i approaches infinity. Therefore the second some is in some sense greater than the former, and Hyperreals are a way of formalizing that quantity.
In practical considerations, and related to why IEEE 754 defines a set of numbers, quite often you need to compute a rough quantity for an expected result. In this kind of analysis you use the same formula that compute the quantity precisely. However, since the calculation is rough, you don't use the whole set of real numbers and real arithmatic. You may treat 10+10 = 10, because it's the same order of magnitude. And you might treat 1/0 as Infinity. The equations are the same, you're doing algebra the same way. The difference is the rules of arithmetic. It's just easier. Making infinity a number makes this kind of math easier too.