Pretty complicated, actually, compared to what would be necessary for primitive silk production from simple pores on the feet. Again, compare the intricate production of a modern orb web (which involves multiple different kinds of silk, as well as precision start/stop control and fancy "architecture") with, say, a slug's slime trail.
In addition, most insects that spin silk or other viscous substances similar to silk do so from abdominal sources. Would this not indicate that this was the originally evolved method and not from the legs?
Maybe, maybe not. If all the "abdominal spinners" are members of a monophyletic group relative to a more polyphyletic grouping of "foot (or other non-abdominal) spinners (i.e., if abdominal spinners all share a more recent common ancestor than all silk spinners of any type), then this would indicate that a) abdominal spinning evolved once, b) all modern abdominal spinners descended from the first one, and c) non-abdominal spinning evolved even earlier than abdominal spinning and abdominal spinning likely evolved as a special case of a more generalized spining ability).
The relative abundance of abdominal spinning might simply be the result of abdominal spinning being a much more effective strategy, and those insect groups which have it have prospered and diversified more successfully than those silk-spinners which didn't achieve this "breakthrough".
Very interesting. Thank you.