What is the difference from the push and pull development models?

Question

I was reading Extreme Programming Explained, Second Edition and in the chapter 11 "The Theory of Constraints" the authors talk about the old and obsolete "push" development model and the XP way, the "pull" development model. It looks like a quite important concept, but it takes only a very small paragraph and two images that are mere illustrations of the "waterfall" and iterative process, nothing specific about these models except by the image caption. I searched and it doesn't go any further about it in the rest of the book. I couldn't find any further explanations or discussions about it in the Internet either.

If the only difference about those is that one is "waterfall" and the other is iterative, them why push and why pull?

Does anyone understand what is really the difference between those two and give some good examples?