I read an excellent article in the New York Times about coming up with methods for teaching the next generation of teachers (or current teachers, for that matter).
The problem is, in a nutshell, that we don't know what makes good teachers good. Economists are happy with this, because they can set up an incentive structure, but as the article points out, teaching is already the 3rd largest profession in the U.S. There aren't a lot of other industries to steal talent from. That means that at some level, the only option is to make the teachers you already have better.
A great deal of the article is devoted to Lemov's Taxonomy. This is mainly the work of one man who set out to categorize the specific actions that good teachers take to maintain control of a class and teach effectively.
Perhaps this is the light at the end of the tunnel. At this point it looks like it needs some larger validation, but the promise of it is that if the taxonomy is correct and effective, you should be able to see results within a year, rather than within a generation.