About time people start thinking like that - assembler hasn't been the smartest choice for quite a few years now considering large projects and maintainance. Maybe for limited parts of large projects that require optimisation and speed but not for whole projects.
Computer Concepts and Sibelius saw this a long time ago, they wanted serious C++ libraries for further development, they weren't delivered and seriously large players left the market. I think most of us missed the significance of those companies leaving RISC OS but that point in time really looks more and more to be the beginning of bleak future to come. Hopefully it can be turned around but currently things aren't looking that great despite a few software and hardware releases.