Odd,
What is the difference then between yours and Armidale's? Because his does wrap to the next page.
(Long-winded reply warning...

)
I think the best way to describe it is to say that Armidale's sheet is less aware of printed pages at all (which is NOT a criticism, btw; it's a perfectly reasonable design choice for anything browser-based), so with his sheet you see some of the same default behavior that you would get if you printed any arbitrarily-long web page... i.e., it gets cut off at the end of the physical page and just picks up where it left off on the next physical page. He's modified that default behavior only slightly in specific areas.
For better or worse, I made a decision early on to attempt to make my sheet printer-page-oriented; i.e., I was going for a WYSIWYG effect between the screen and the printed page, and I took advantage of the latest HTML5 features for controlling printed layout.
That decision of mine has had unforseen consequences. It worked great at the beginning, when I was doing everything for my own benefit, on the browser I happen to use. Then, as more people were interested in the sheet and I added support for other browsers, I discovered that the print-related features of HTML5 are not fully supported yet (or are extremely buggy) on certain browser/OS combinations. So, I had to compromise and back away from using those features in certain areas.
The result is what you see today. It works in most cases and does have some advantages, but it "fails" (from a layout perspective) in this particular situation.
Why haven't I ditched this approach altogether, then? Because it's still my goal to achieve my original design and make it WYSIWYG. I'm keeping track of the relevant defects in the bug databases of the major browsers so that I can do that when they release their fixes.