I think you're basically right, but I don't think things are necessarily completely shot for RISC OS. Although we may never get another ArtWorks written from scratch, that's partly because in some areas key RISC OS software is mature and does a good job. And there are still applications being written which provide useful functionality: think of Parmesan in the last few months.
Don't get me wrong, RISC OS has a lot of challenges and may not be able to meet them. But there is work being done or planned on the OS's major deficiencies. If the ROOL team can get the full OS into shared-source, and then ported on to modern, portable hardware, who knows? We may just be able to maintain a viable, hobbyist platform and re-kindle some enthusiasm for new application development.