Good thread (I think).
I've read of (but not experienced) .45ACP that fires based on the extractor, and with excess headspace. Reliability is questionable. The example was .45ACP fired in a .460 Rowland chamber.
I can see that, without the shoulder to provide a positive stop, and with this huge amount of headspace (essentially in a barrel chambered for a longer case), the extractor may just not be up to the task.
So the ignition reliability myth is just that, and the main concern is enough headspace for proper chambering (especially in gas guns), but not enough to shorten case life or risk case head separation.
As to the question about "head clearance", I believe that is another expression for proper headspace. If you have zero head clearance, you may be just fine for a bolt-action rifle (which has tremendous camming power when chambering) but not so good for a semi-auto rifle (which does not.)
In other words, a certain amount of head clearance is necessary for ease of chambering. Depending on the action, this can be critical. (Look up "slamfire".)
I haven't built an AR-15, but I understand they are essentially drop-in if all the parts are mil-spec. Here's a link -
http://www.ar15.com/mobile/topic.html?b=3&f=4&t=315921