Very nice thread !
Thank you, [MENTION=59082]Mercurius[/MENTION], for opening it. Paying hommage to Boorman's Excalibur and Moorcock style was just icing on top of the cake
If I read well the excellent contributions of [MENTION=6696971]Manbearcat[/MENTION] and, especially, [MENTION=42582]pemerton[/MENTION], I feel there is very little difference between "old" and "new" school. I am more worried with what's in between...
Let's dive into it : I fully endorse the claim that there is a very special imaginative process at work at the heart of "Fortune in the Middle" mechanics. Particularly, you have strong constraints on where you are coming from, and where you are getting to, and a lot of freedom concerning the journey from A to B. This architecture provides many features, as it provides some room between gamining mechanics (which can focus on fun, balance, ease of use, whatever) from narrative license, and empower each player with his own vision of "what happens" while guaranteeing everybody's on the same page concerning the final state. Of course, I have seen many people, on this board and elsewhere, disliking this kind of mechanics, and putting the blame on 4E for using them so casually... I feel they are misrepresenting the issue, as the most blatant supporter of FitM was, IMHO, Gygax himself, especially in 1E DMG. When you think about it, with his random tables, the old man invented the Schrödinger dungeon ! The room behind the door doesn't have to exist before you open the door ! This is the ultimate scene framing device, not in the sense of creating meaningful scenes, but in the sense that only the present "encounter" matters. So, no, I wouldn't oppose Old and New schools on the way they use and support imagination, on both a microscopic (rolls are FitM) and macroscopic ("scene framing") level.
I am under the impression Next is pretty neutral concerning FitM mechanics : it uses them quite a lot (in fact, once again, HP are going to be a real source of headaches for those who don't embrace the FitM paradigm), but not as overtly as 1E (I would say 4E is not blatant enough on this subject, and this is one of the many reasons of its demise). Next also promotes a time framework compatible with scene framing, with short and long rests recharging PC resources. Where Next does an excellent design job is when it tries to enable this protagonistic time management device while preserving the naturalistic flow of time. This philosophy shows at many places, where the design tries to make room for "kool powarz" without the (IMO) obnoxious power formatting (I hope Next Fighters will be able to lure foes into battle, spending some metagame resource, but not in the cold and tokenized manner of "Come and Get it !" encounter power). I also feel that the design team believes (like me, and [MENTION=42582]pemerton[/MENTION], I guess ;-), and I am sure the class has a special meaning to [MENTION=697]mearls[/MENTION], if I remember Iron Heroes well enough) that the game vanquishes or perishes with its implementation of the Fighter, and they are definitely aiming at giving it back its "Old school" resiliency.
For those afraid to lose 4E "story now" features with Next, I would say I think there is more to gain than to lose with the relaxing of the hygienic regimenting of time, resources, roles, and threats. (For instance, I don't feel able to DM a Song of Ice and Fire style campaign, with a naturalistic take on the world and the idea that every character can be a protagonist, using the 4E framework. Having to rebuild NPC when the PC level up, or when they acquire PC status, seem very cumbersome to me). A less tokenized system, besides being less alienating to a LOT of players, would open up nice possibilities such as the Doom Pool : my real concern, regarding DM force, has always been restraint : why shouldn't I Buff, Scry and Teleport to get rid of these nasty PCs in their sleep ?
I definitely think 4E "tight math" and p42 are overrated : math is surely better than in 3E, thanks to the enforcement of a "bounded accuracy" policy, but the ubiquitous use of a single die (a d20, for instance) as a randomizer yields to a very strange world indeed (take a closer look at the distance jumped, for instance...). Actually, "math" was tighter in 1E, because Gygax understood the use of a bell curve, and used one (or more !) new table for each and every case he encountered. I can see the appeal of a unified mechanic, but 4E is really : you have 30% / 55 % / 80 % to succeed at a Hard/Moderate/Easy task appropriate to your level. Adjust by 5% for every 2 level difference. Even Fate (which is not what I would call a crunch heavy game !) has a richer base system, as it starts from a bell curve... Also, all those minigames contained inside D&D have always fired my imagination : how many armies built with the War Machine ? Castle features carefully paid for ? Decks of many things and Wands of wonder and Spheres of annihilation and Vorpal swords ? Psi using BBEG ? A clean system is serviceable, but it sometimes comes to the detriment of accuracy (jump length), effect (fighters resilience), fun & imagination...
(by the way, [MENTION=42582]pemerton[/MENTION], I think you are overrating/overselling 4E a bit too much with the support of inspirational healing. In Tolkien and Arthurian Romance, both (master)pieces of reactionary literature, healing IS divine : both Arthur and Aragorn are king by divine right, and I would argue their ability to *channel* heroism and bravery in their followers, or to be themselves inspired by an icon of love, is divine by nature. So, IMO, these characters ARE paladins, and divine healing is inspirational by nature. To be continued

)