Wow, I should have checked in an hour ago.
Okay, so, brief comments on a variety of responses so far:
Luffing when wind "in line": I'm pretty sure the wind is "in line" with the luff of my sail when the telltales stream evenly aft on both sides. The sail isn't luffing because it's curved. The curve is maintained by the angled pull of the sheet.
Lift: perpendicular to freestream flow, not to the foil. Minor point; you could break down the force of the wind into any components you like, but only one is referred to as "lift" in the textbooks.
Acceleration on the lee side: my understanding is that faster flow separates sooner, which greatly increases drag, and that therefore you actually want to decelerate
flow on the leeward side (while still keeping it faster than the windward flow), and that this is in fact how the slot effect works.
Originally Posted by asdf38
(paraphrased) The lift model lacks in quantification
No, that's patently false. Lift is given by a integral of forces over the surface of the sail. It's not especially useful for internet forum discussions, but it's there.
So yes, your "propeller pitch" model has an easy quantification --- which, as I keep complaining, is horribly wrong, because it predicts infinite speeds. Oh, except that you add drag, which you don't quantify. So your model either predicts infinite speeds, or is not quantitative, depending on which features you include :P
Here is a new point - think about how neat it would be if you could sail off apparent wind.
But, you do sail off apparent wind. If the apparent wind is in a direction that permits sailing, that is. Same goes for true wind.