Skeletron:
Well…, you have to physically get the drilling or punching equipment in there to make the web hole; some nominal size washer has to clear the flg./web radius, the “k” value, but I have clipped the o.d. on some washers to make them fit in special situations; the steel manuals used to show a g1 value, as in your method A.; and finally you have to be able to get wrenches in there for erection. We are often most interested in the relative top of stl. elevs. The biggest reason probably pertains to layout and detailing, where we have std. starting surface, and today’s shop equip. just presses the top surface up against a stop (ref. point) and goes from there. The web center is kinda a nebulous location for the shop layout guys. Get out your calipers and micrometers and tell them where the center line is.

Dhengr: I appreciate your input. And you nailed it (bolted it?) right on the head! You need to provide a reference "zero" for your machine. Top flange is the only way I see it. The office detailer armed with 10 fingers, the OFFSET command, and midpoint snap turned on, doesn't see it that way...

Your office detailer can easily align the bolts more or less centered on the web depth and still dimension them from the flange, right? Any eccentricity due to rounding would be negligible.

HotRod10: Yeah. They still give the dimension from top flange. The difference is negligible and even more meaningless on a small fab job (<20 pieces, 1-2 beam sizes) compared to a big industrial project (large tonnage, multiple beam sizes, etc.). All these observations (specifically in this forum topic) are really funny to me because in the fabrication world (as a connection designer) we gave a lot of energy to these details. But as a design engineer, specifically on smaller jobs, the focus doesn't get to that level.