Liam,
I won't respond point by point because I agree with most of what you're saying. There has been, for the last 3-5 years a real need for GNUstep to start a "reference distribution". I think it was discussed here on the list more than once. The issue has not been the desire to create one, but the logistics of doing so. I am thankful for the existence of GSDE as it is very close to what I had imagined as a GS desktop. Originally, the push to make GNUstep an API only thing was my idea... I believe that the focus on getting the UI done has helped us in lots of ways, but has, perhaps, hindered us in others. We focused on being more cross platform rather than JUST on the one environment we support, so it has had some benefits. But now, it's time for us to create a DE, perhaps in collaboration or as an enhancement to GSDE to show EVERYTHING that GNUstep is capable of.
There are a number of things that need to be refined in GNUstep:
1) Printing -- printing is in an AWFUL state... it doesn't produce acceptable output in my opinion.
2) I had started an effort to get us to parity with APIs on Catalina... there is still work to do on this... even though some people believe we still have work to do on being OPENSTEP compatible (Riccardo, I would love to see what you believe we are missing).
I believe being a cross platform development API has benefitted us... but I also believe that we need to appeal to the masses as you have pointed out.
The only issue I had with your article was touting OPENSTEP or NEXTSTEP... my objection to the use of those 30 year old terms seems as strong as yours to my use of Cocoa which I still don't believe is deprecated. That retired document shows BlueBox and Carbon I believe... those are likely what is being deprecated, not the Cocoa name, but whatever. :)
Yours, GC