And then finally e-collar use (positive punishment).
I think the lines are blurry here between R- (negative reinforcement) and P+ (positive punishment). Teaching recall with an e-collar can be done more than one way. My preference would be R- and the method would be to use a long line, let the dog wander around, give him a working level (lowest setting at which he can notice the collar) stim, hold it and draw him back to you with the leash, release the stim as he comes to you. In my experience, within ten minutes you have a dog that has learned to "turn off" the mildly annoying sensation by coming to you when it is felt. Once this response is solid, you name it with your command word and issue the command before using the stim. Soon no e-collar use is necessary most of the time. Then you can add distractions. At this point there has only been R- and not P+. (R- increases the likelihood of the desired behaviour, reinforcing it, in this case the desired behaviour is coming when called)..
Using the collar to break up a chase or dog running away from you when you have called him would be an example of P+ to stop the unwanted behaviour. In this case you would use a high level stim (P+ decreases the likelihood of the unwanted behaviour, in this case chasing or running away) and immediately revert to a whistle or vocal command (or working level stim if needed) as soon as the dog stops his chasing/running away.
I know the lines between R- and P+ can be blurry, but the proof is in the results. If it increases the likelihood of the behaviour it is R- and if it decreases the likelihood of the behaviour it is P+.