This is shaping up to be an exciting development. It could still fail due to an unworkable license or a refusal to release the most important components but if this project can deliver what it promises, this could push RISCOS forward.
I'm glad that Castle are working on a restrictive license; it's a workable strategy as opposed to 'let's just give our products away'. Any commentary is purely speculative until we get a finished license and an actual release of code.
I agree with JohnB that Castle's proposed strategy offers benefits for all parties involved.