Doran and Lasenby use bivectors with a minkowski -+++ metric:
\sigma_i = \gamma_i \gamma_0 as a spatial basis. These have a positive square, like a set of euclidian orthonormal vectors, and thus behave vector like for all intents and purposes.
They write the electromagnetic field as the following bivector:
<br />
F = \sum E^i \sigma_i + I B^i \sigma_i<br />
So, one can recover the electric and magnetic field components with:
<br />
E^i = F \cdot \sigma_i<br />
<br />
B^i = (-I F) \cdot \sigma_i<br />
Other than being a cool way to condense maxwell's equations into one equation, I haven't gotten as far as figuring out how to actually utilize all this in calculations. I've had to go back to basics and understand SR fundamentals a lot better before I can understand their applications of GA to SR, and E&M. The two days that we did SR in first year E&M back in my undergrad engineering days wasn't enough to to able to understand chapter 5 or 7 of that book;) I'm starting to get a handle on some of the SR basics but still haven't gotten back to the GA side of things.
fwiw, I've collected some notes on both the Hestenes and D/L texts. Both are too dense for me and I found I had to spend a lot of time mulling over details to be able to understand things sufficiently. I recently organized these notes a bit and dumped them here:
http://www.geocities.com/peeter_joot/
Perhaps some of it would be of it would be of interest to you ... I wrote it up to explain things to myself, because if I couldn't do that I found I didn't understand something.