I think what many of us are pointing out is that taking death off the table has somehow become common wisdom
I'm not really sure who you're speaking to here.
No one in this thread has advocated taking death of the table. (In fact, [MENTION=85870]innerdude[/MENTION] made a point of explaining that it remains on the table in Savage Worlds.)
This whole digression onto character death arose because innerdude and I were comparing thoughts on how 4e and Savage Worlds, via their PC build mechanics + action resolution mechanics, push players to take their PCs into unanticipated sorts of places in the fiction - heroics when the player intended a non-heroic PC, or holding the front line when the PC is far from heavy infantry in capabilities, or a player throwing his/her CHA-dump PC with few social skills into the thick of a social situation because s/he doesn't like the direction the "face" PCs are taking it.
In that context, I said (post 581) in relation to the latter two sorts of occurrence that:
This is not necessarily where heroism comes to the fore, but it can lead to narrow and exciting successes, or to meaningful failures - meaningful because the player really put his/her PC on the line for a reason, and even if it didn't work out quite as desired something interesting happened in the fiction as a result.
I think 4e supports this better than AD&D because its approach to action resolution, consequences etc creates a more nuanced range of failure and partial failure conditions than simply alive/dead.
My comment has nothing to do with whether or not PC death should be on or off the table, or makes for a good stake. On the latter, I think [MENTION=90370]Zak S[/MENTION] has given a good account of the power of PC death as a stake in a certain sort of game (from the outside, it sounds to me like a classic "skilled play" dungeon exploration game of the sort that Gygaxian AD&D and Moldvay Basic are aimed at - if I've got that wrong, I'm happy to be corrected). I've tried to explain why, in the sort of game that I like to play and run (from the inside, I would describe it as a game in which player identification with the PC is similar to the vicarious emotional experience to which dramatic fiction gives rise), PC death is not necessarily the ultimate or most powerful stake.
Going back to what innerdude and I were discussing: in a system in which the main, or at least ultimate, consequence of failed action resolution is PC death, two things follow. First, a player always has a reason to try and engage the ingame situation using his/her PC's best ability; second, if a player (due to rationing, or misadventure, or whatever) finds him-/herself engaging the ingame situation using a PC's weak ability, something has probably gone wrong. The upshot of these two things is that skilled players try hard to avoid (or at least manage the occurrence of) situations in which those weak abilities have to be used.
For instance, in Moldvay Basic if a MU PC ends up in melee, something has probably gone wrong. Good player recognise that sometimes even the MU might end up in melee, when all else fails, but they try to avoid those situations. And when they realise that such a situation can't be avoided, they try to manage it as sensibly as they can.
In a system in which failed action resolution doesn't necessarily mean PC death, or getting closer to PC death, then even if it remains true that a player always has a reason to try and bring his/her PC's best ability to bear (which generally remains the case in 4e), it doesn't follow that if some weaker ability is being used then things have gone wrong. Once outcomes other than PC death are on the table and highly salient, the reason a player has to deploy his/her PC's best ability can easily be overridden by other reasons to do with the fictional situation and the possible outcomes in the fiction. (Those other reasons will only arise if the player is sufficiently emotionally invested in the fiction to care about that range of outcomes. Hence my comparison, upthread, to dramatic fiction.)
It's possible for a system to go even further in this second direction by taking away the reason that players have to always bring their best abilities to bear. Burning Wheel does that, because its advancement system requires players to take on tasks that they can't succeed at - so a player has a reason not to always use a skill at the maximum bonus it could be used at.
[MENTION=85870]innerdude[/MENTION], how is Savage Worlds in this respect - do players ever have a reason to avoid bringing their full bonus to bear even if they could?