Aug. 27th, 2013 edit:
Tapping the output transformer as a balanced secondary is totally goofy as was determined by Osvaldo de Banfield and Cnpope (Chris) here:
I'll leave it posted here just to show that I'll try just about anything and to show that it's unwise.
Aug. 7, 2013 edit:
I was just looking for the Hawksford and Mills current amp paper on google and I found this very blog post so I thought I would point out a potential drawback to the following feedback technique. Hawksford and Mills claim that using a sense resistor to the ground return from the speaker introduces a distortion similar to "interface distortion" in voltage amplifiers.
There is very little I could find about interface distortion.
As far as I can see it, this feedback scheme may introduce back emf from the speaker and feed it back to the amp...