That alternative is the books published or republished by Dover publications. They like to take older textbooks and purchase rights to republish them as relatively inexpensive paperback editions. A very large fraction of their books are under $20, with many under $12. A few are more expensive, but only rarely more than $30.
The level ranges from suitable for high school students to graduate level and beyond.
Here's their mathematics section: http://store.doverpublications.com/by-subject-mathematics.ht...
Don't overlook the "general" subcategory. They have some wonderful problem books there, such as Yaglom and Yaglom's "Challenging Mathematical Problems With Elementary Solutions" series.
They also do this for physics, chemistry, engineering, history, economics, computer science, biology, earth science and more.
Hey IEEE, you are doing the opposite of a service in a world created by your members. Please cease to exist.
As an on-again, off-again IEEE member, I've often wondered what it would take to stage a "revolt" of sort and get enough like minded people on the IEEE board (or whatever they call their leadership group) to radically remake their approach to monetizing content. In an ideal world, I'd love to see all research papers made freely available and the organization funded solely by membership fees and other means - conference fees, sales of dead-tree books, probably other stuff. Maybe charitable donations?
But up to this point, I'm afraid I've never been motivated enough to really pursue the issue. But if anybody wants to form an "open access caucus" (or if there already is one!) feel free to give me a shout. I'd still like to help with an effort of this nature.