Removing from one side then the other. I've been free sharpening for a year now and I know all the basics and the concepts. It's just more of technique advice that I could use. On another note is it ever advisable to use trailing strokes on the stone to finish off the knife? I tried to do that on my glass stones and I get in an endless cycle of flipping the burr.
I avoid edge-trailing strokes at the final stages, because I think they produce wire edges and chasing-burrs as you mentioned. Beyond that, if you ever read verhoven's experiments (
http://www.bushcraftuk.com/downloads/pdf/knifeshexps.pdf) it does a good job of explaining why edge-leading strokes remove a burr so much better.
As far as technique goes I'll just describe what works for me... Once you're at that point where you really
know that burr is gone from both sides and you start doing alternating strokes from side to side in your final "honing", maybe throw in a little extra attention per side say two or three strokes and not with such great concern with light pressure, but with contact of the apex on the abrasive. I think that in the proccesses of removing a burr, the apex still needs to be refined/established back a bit at the point where the burr was broken off. If you just do one stroke, alternate to the other side, do another stroke and at extremely light pressure I think it just tends to take a bit longer and therefore makes you more prone to shake or screw the angle up, or never actually get the apex to its true optimal point without going over, or degrading it from the burr removal. You're better off after having removed the burr to spend a little extra time and pressure on each side--minding that you don't raise another burr--before alternating side to side with the
extremely light pressures. At this stage I really go mostly by the feedback the edge gives me on the abrasive to tell me when I've got that apex fully back with no burr before the final light-pressure honing. Might sound like I'm talking about the technical aspects, but really what I'm saying is that once you're at that point of edge refinement, human error is as such that one pass on one side and you may have not actually hit the apex--or you may have overshot it. There's a certain level of error in my own angle contorl I learn to compensate for in that way, and by doing one or two passes and
then alternating it's kind of like suring up my odds that on at least one of those strokes I got a perfect pass on the apex without undershooting or overshooting it. I might be overcomplicating it, but it's hard to describe technique...
Some people do encourage edge-trailing because when you remove the burr, it dosn't
tear it off and leave this blunted portion that still needs extra attention to refine back, but I think it tends to just raise a burr out straight and leave a wire edge.
Most of this stuff does ultimately come down to personal preference and what works, there's no real "One way works, this way doesn't". I'm just a fairly strong advocate for
ultra light, edge-leading strokes to do the final finishing/honing.
To give an idea of the type of pressure I'm talking about, take a piece of tape and put it between two books, not too tight, not too slack. Then touch your finger to it without making it sag at all...