Ultimately, the GNUstep project should have taken on the burden of porting whatever could be kept from Windowmaker into the GNUstep world, and made a complete solution to compete with GNOME / KDE with a chance of also making it easier to write an app for both OS X and Linux using much of the same underpinnings (Objective-C, etc).
A lot of really great work went into GNUstep but looking back... for what? It would have needed so many more resources than it had to ever have a chance at becoming something like the above.
Maybe with AI, one day it will be easy enough to pick some of these ideas back up and run with them.
We are not really in a position to dictate what the GNUstep project ought to do. I for one am rather thankful for what they have achieved so far. Especially since they seem to mainly lack the man power which its critics probably won't provide either.
Not being a software developer myself, I've resorted to loosely following the progress of GNUstep over more than 20 years, in the hope that it will eventually become a viable option. But apparently capable people found more interest in KDE, GNOME, and similar projects based on either Qt or gtk+, and then hardly anyone cared anymore.
By all means, if anyone here is capable enough is wanting to still contribute to the GNUstep project, please do! At the current stage of the project it would be a pity if all the efforts made over the years would be wasted by not supporting it anymore.
Well, GNOME it's becoming RedhatWare disallowing Adwaita theming and forbidding custom vendor icons. All when Gnome 2 with Tango (and before, BlueCurve) were one of the best designs ever. Readable icons with outlines, clear themes, no bullshit, and a very lightweight classic desktop.
On KDE, hope QT and Nokia or whoever controls QT today doesn't crap out everyone with a new licensing again. Because if that happens, Plasma it's dead. And Marble and tons of nice projects such as Okular. Yes, QT forks, but that takes tons of time.
A lot of really great work went into GNUstep but looking back... for what? It would have needed so many more resources than it had to ever have a chance at becoming something like the above.
Maybe with AI, one day it will be easy enough to pick some of these ideas back up and run with them.