OpenOffice's website design, marketing efforts, and documentation are less friendly to non-technical people that those of LibreOffice, the most prominent fork of OpenOffice. Subtle "superficial" details matter a lot to non-technical people -- for example, LibreOffice's web page has a "Get Help" link, instead of a "Support" page.
In addition, LibreOffice has been the default office suite in key GNU/Linux distributions like Ubuntu for many years.
As a result, the LibreOffice fork has been more widely adopted than OpenOffice, leading to the present situation.
One possible path for the OpenOffice folks would be to copy the Linux kernel and become an "upstream project" that others use to create a range of office-suite distributions like LibreOffice. In other words, stop packaging and marketing OpenOffice to end-users. Stop competing for end-users with downstream projects.
"One possible solution would be for the OpenOffice folks to copy the Linux kernel and become an "upstream project" that others use to create a range of office-suite distributions like LibreOffice."
... which were exactly the thoughts and reasons behind Apache accepting OO in the 1st place. For various reasons this obviously never happened. But it WAS the idea from the start.
Interestingly, that could have happened. Not only did LO make the effort of rebasing all their changes on top of a particular AOO release, they also were set up to incorporate all the new work from AOO into their branch. For instance, they quickly incorporated the sidebar code. Even today, one of the LO developers still goes over each AOO change, merging any worthwhile ones.
What actually happened? For some reason, development on AOO stalled. What little work still happens on it, is almost always replicating work LO already did two, three, sometimes five years ago. See it for yourself at https://cgit.freedesktop.org/libreoffice/core/log/?h=aoo/tru... : pick a commit, and look for the "Notes" field. For instance, picking "Port main/fileaccess to gbuild" I see the note "prefer: 1f9bc2b2aa8168f9c164044058b117d2a17d83ad"; that git hash is for a LO commit from 2011 (https://cgit.freedesktop.org/libreoffice/core/commit/?h=aoo/...).
In the meantime, LO did a lot of cleanup work. Finishing the build system move to gbuild, cleaning warnings across all the code base, removing plenty of obsolete or unused code, running several static checkers, and more. As I noted at https://lwn.net/Articles/699108/, the recent AOO CVE was probably found and fixed on LO two years ago, as a result of running a fuzzer together with valgrind. If you wanted to pick an "upstream project" to base an office suite on today, it makes more sense to pick LO instead of AOO.
I think you are proving my point... AOO was "plundered" by LO but LO never gave anything back. So any "co-opporation" was extremely one-sided and one-direction. It was hoped that LO, as good FOSS neighbors, would donate back, even though they didn't have to, just as we hope that our corporate neighbors also donate back. We trust in altruism rather than in forcing it. It just so happens in this case that our trust was somewhat misguided.
In addition, LibreOffice has been the default office suite in key GNU/Linux distributions like Ubuntu for many years.
As a result, the LibreOffice fork has been more widely adopted than OpenOffice, leading to the present situation.
One possible path for the OpenOffice folks would be to copy the Linux kernel and become an "upstream project" that others use to create a range of office-suite distributions like LibreOffice. In other words, stop packaging and marketing OpenOffice to end-users. Stop competing for end-users with downstream projects.