Actually, at least two of those (E and WindowMaker) remain alive. And WindowMaker's got its fervent fans (you're hearing from one here).
While Raster's continued to plink away at Enlightenment, among the reasons WindowMaker development's been so modest is that it accomplished its mission: provide an implementation of the NexTstep interface. I use wmaker without most of the rest of the GNUstep tools (I find them kind of funky and cumbersome), but the window manager itself is simple, straightforward, and rocks.
It's also very similar under the hood to Aqua as used now in OS X, which for the most part just skins it differently and removes a bunch of features I like -- so while I love wmaker, I really can't function on Macs.
As for userbase. Yeah. I'm aware that I'm in the minority. I'm totally OK with that.
BTW, putting off window manager/desktop environment philosophy arguments, Input Method Engine is one of those constantly neglected aspects that really matter for East Asian users. It seems that the ones making plans for WM/DE and other infrastructure had little overlap with users, and their designing decisions were very likely to omit the requirements necessary to cooperate with IMEs.
I'm not enough of a dev to know what would be required, but pitching this to the developer(s) might be helpful.
The only thing that's not happening is a whole lot of development. And I'm actually totally OK with that.
>enlightenment - okay, kind've resembles Win 95 too... but more of a DE than the others you listed
> etoileos - last news update 2012 and wasn't even about the project
What you seem to be missing is that these projects died because people thought they were weird, which is to say, innovation in the DE space pays negative rent, and that's no good, especially not from the perspective of an open-source project, which, the social dynamics of open-source dictate that projects need to acquire a large userbase to sustain an active development community for more than a couple years, that is to say, to make people keep working on it after the "new project smell" wears off.
In other words your observation is a direct consequence of the choices in DE that users have made and continue to make.
The recent trend has been towards modularity, and while you dismiss awesome and ratpoison, a major boon of LXDE et al is that, unlike Windows 95, you can replace the window manager with xmonad and still use all of the other components of LXDE. Modularity brings innovation to the people who want it while satisfying the large majority of users who apparently do not.
There's no reason for LXDE to ship anything but Openbox; LXDE could certainly switch to xmonad tomorrow, but their users wouldn't be happy. And who wants that?