I consider Wayland's choices reasonable as well. I.e. I it's not surprising that a reasonable attempt to clean up X ends up looking similar to Wayland, just in a slightly different place on the xy graph that has backwards compat and cleanup as its axis.

The thing is, X11/Xorg is a huge monolith. There is simply no way to implement every single feature in one go and then release the competing implementation.

The very thing that makes people biased towards X11/Xorg both negatively and positively is that it is a huge monolith and the only X implementation on Linux. The moment you have two implementations, you're gonna get the same complaints against the second X server as Wayland is receiving.

You think this is an indictment to the second implementation and that they shouldn't bother by saying an "analysis was incomplete" but in my opinion it's exactly backwards. This is an argument that eternally perpetuates X11 not because of technical capability but rather because it was there first.

After all, the moment there is any implementation that is second it might miss a single feature that nobody actually uses, but theoretically could be used when combining an old binary with a new X implementation.

But this argument misses the obvious fact that X11/Xorg is already dead since any code change will break existing applications. Meaning X11 has become an unsalvageable fossil.

The Linux kernel is also a monolith yet has worked forever without a rewrite.

I think there is a reasonably agreed on set of things that can be removed from X, like server-side drawing primitives, and GLX

The Linux kernel was not born obsolete on arrival like the X Window System was. At least, not fully obsolete.

It will be obsolete someday, anyways.

“Obsolete” means that a replacement exists which is as good as, or better than, the thing it aims to replace, in every significant way.

X11 is far from obsolete.

In what way was X11 obsolete in 1987?