I think HR can be useful, but religious HR training has a lot of problems. One of the biggest is cardiac drift. HR afficianados claim that well trained runners don't show cardiac drift, but that's simply not true, especially if the temperature is at all above optimal. It's just that well trained runners show less drift than poorly trained runners. At a given intensity, HR will steadily climb throughout a workout, so trying to stay in a particular zone will usually lead to an athlete slowing down.
I prefer to train by pace and perceived exertion, using HR for post workout analysis, or in rare cases to confirm that I need to pull the plug on a workout that is going really poorly.
For example, if I'm repeating a workout I did three weeks earlier, I can compare my HR curves. If I went faster at the same perceived exertion more recently, but my HR curves are higher, that probably indicates that I'm approaching a peak. I'm likely using more of my anaerobic system to sustain the pace and I've also probably trained my central governor to disregard some of the pain I experienced three weeks earlier. I know that continued progress along this line will probably be limited.
On the other hand, if I do the same workout faster but with lower HR curves, that probably indicates an increased lactate threshold. It also likely means that I'm still a good ways from peaking.