Are you going to be recording to a computer / workstation?
If so , it is really easy, although it does involve a little trial and error.
You point your fig 8 mic with the pickup pattern side to side to the source you are recording (the null point facing direct at the source). You place the mid mic (generally in cardiod pattern) as close to the diaphram of the fig 8 as feasable obviously pointed directly at the source.
You record the mid mic to one channel and the fig 8 to another.
You duplicate the fig 8 track to a second track (make a clone) but invert the phase of it. You then pan the 2 fig 8 tracks opposite each other to taste.
Bring up the center channel first then the sides, if you hear some phase cancellation going on, reverse the phase of the 2 side tracks. This is the trial and error I was talking about.
What is happening is this. The fig 8 mic is picking up information to the left of the source that has (let's say) a positive phase and also information from the same source from the right side of the mic that is 180 degrees out of phase.
By duplicating the tracks and reversing the phase of one, you now have the right side track where the material has a positive phase on what was to the right ( but originally negative) and a negative phase on what was to the left(but originally positive). When panned out wide the left and right room information both have a strong positive phase, and all the info going to the center cancels itself out making a dead hole. This is where the center mic comes in, to fill the hole.
The closer you pan your outside tracks you may run into a little weirdness. but the M/S configuration makes for a really big stereo field, which works particularly well if the room sounds good.
This is probably a pretty lame explaination, but is the way I understand it. I a sure others will come along to correct my technical mistakes.
Tom