Brad DeLong graciously linked to my post extending/critiquing Steve Horwitz's jigsaw puzzle metaphor for the difference between Austrians and Keynesians, which I felt relied on critiquing a parody of Keynesianism.
It's fitting that Brad linked to it because I think he's one of the best guys out there for explaining that Keynesianism is not about the government being the spender of last resort - it's about the government acting to eliminate distortions in the interest rate. These distortions arise from the fact that the interest rate is really one price for (at least) two goods: credit (or loanable funds) and cash (or liquidity). For those of you that remember your linear algebra, this is the same as saying that we have more equations than we have unknowns. We have four equations (demand for loanable funds, supply of loanable funds, demand for cash, supply of cash) but only three unknowns (quantity of loanable funds, quantity of cash, and the interest rate). We call systems like this "over-identified", which means somewhere there is going to be slack. Keynes was ambiguous about where the slack would come up. Hicks forced the slack to express itself in output by making a model where both the loanable funds market and the cash market had to clear. Wherever you fall (and I'm personally agnostic between Keynes and Hicks), the point is it's this price distortion that causes all the trouble, and Keynesianism is about fixing broken interest rates.
For some reason, this message does not get across to people and Keynesianism instead gets treated like it's saying "Y=C+I+G so if you increase G you have to increase Y". This view reads Keynes writing about burying bank notes and digging ditches and assumes he's talking about make-work projects. This view reads Brad DeLong saying that anything that adds to the deficit passes the cost-benefit test and thinks he's saying we just need to add more demand and pump up G. This completely misses the point. Keynesian ditch-digging is not a make-work argument. Keynesian ditch-digging is a helicopter drop, not a make-work program. The first helicopter flew in June, 1936* - only a few months after Keynes published the General Theory. It's a shame, because if Keynes had come up with Friedman's "helicopter drop" analogy rather than a ditch-digging analogy it probably would have lead to a lot less confusion.
If you want to cling to a loanable funds theory of the interest rate, by all means do your best at defending that view. It's going to insure that you happily and naively live in an abstract world of full employment and supply creating its own demand. Good luck with that one. But if you choose to engage Keynesianism, make sure you take the time to understand the argument. It's not about make-work, and it's not about replacing what the private sector used to do. It's about correcting price distortions that are inherent in monetary economies.
* - In Nazi Germany, which is why the helicopter picture in the top-left has a swastika on it... I don't just randomly post Nazi military images on my blog - it ain't my fault they figured the helicopter out first.