in terms of DM-work, it is far easier to give out 300XP per CR divided by # of players than it is to record how much damage was taken/recieved per encounter.
XP-for-damage-taken can drive an undesired/nonsensical playstyle (people should generally want to avoid getting hurt)
XP-for killing-stuff being the standard, the side-effect is kill-happy PCs, but at least they are still within the concept of people who face danger and kill stuff
Thus, I think for most GMs, XP-for-damage-taken is too much bother at the risk of not making sense anway.
One model I might consider is to determine what the PC's objective is (either a quest, or player chosen goal like "take over the kingdom"). Then look at the straight line path of encounters to achieve that, and add up the XP the usual way. Let's say the party chooses "rescue the princess" and you determine there's 5000XP if they go straight to it through 5 encounters.
Now, let them do the adventure. Don't track encounter XP. If the PCs face 20 encounters because they do some wasteful stuff, or find a back way clever solution, give them the 5000XP. In short, the PCs are rewarded for achieving their goal. The value in any encounters is whatever they get out of it (treasure), rather than XP directly.
If it was a major quest, I suppose you could break it up by stages. Whatever.
Possible side effects:
players would try to skip some encounters, knowing they aren't losing XP
players would still look for encounters, knowing that they'll get loot and reduce a future threat against them,ultimately stuff gets done in encounters
learning from failure isn't modeled in this method, there are valid times the PCs fail and should get XP anyway
learning from failure is a complex issue. In a way, that's what the OP is talking about. When you take damage, you have failed to block, dodge or defend yourself. The question is, is this something a PC would realistically learn from?
I don't doubt a person can learn from failure, but in my experience, failure is not the best teacher.
I find that when I have people who are stuck, or behind, they need some successes to help them catch up and get the hang of things. That usually comes in the form of easier sub-components of the problem.
I find that when I am teaching a complex technical concept, having technical issues during the demo/explanation gets in the way of talking about the actual point.
I find that, in failing at something, we don't always walk away with the same conclusion/lesson that somebody who succeeded did. In many ways, we end up with a kneejerk, more cautious attitude about the lesson which doesn't help us actually succeed and win the next time it comes up.
When they say, "whatever doesn't kill you makes you stronger", they're not talking about something that beats the snot out of you and leaves you for dead. They're talking about something that TRIED to beat the snot out of you and you WON. or they're talking about how after getting the snot beat out of you,you defied the doctor's expectations and WON. The learning comes from the process of winning, not the losing part.