Why Rebol was never open-sourced I will never know. I guess Carl came from a different time and the non-Unix/non-academic world where software tools were sold at a high price.
Rebol has so many interesting ideas in it that it really could have carved a respectable niche in the world of networked programming languages. Just having sensible pragmatic built-in types for email and money was enough to bring it to my attention.
It's is a very odd but amazingly powerful little language worthy of study by anyone interested in language design. It has been a big influence on me even if I have never used it for anything 'real'.
It may be the case that Red was created for the reason you state, but it simply doesn't appear to have the necessary momentum. And that's unfortunate, because REBOL does not have a lot of interesting and smart features, such as having mini-DSLs ("dialects") for things like GUIs.
To date, Red can't even run without already having REBOL installed on your system. The project has been around for almost two years now, and only a few days ago has it successfully been able to do the program "print 1".
Red was announced and started last year in March. It's basically 2 languages and until last month development was going towards compiled low-level dialect Red/System. I was actually impressed by speed of things moving in Red/System and many members of small REBOL community contributed to it. (It had GTK bindings, ZMQ, port to android, while it was still in development).
The announcement this week about "print 1" is about the fact that now the "Red" (higher, rebol-like level) language is also at the brink of working.
what really, really impressed me about rebol was its runtime. it seemed to achieve an almost magical combination of small, cross-platform and capable, bypassing the usual "pick any two". but even back in 1999, when i first came across it, i couldn't bring myself to invest in a closed-source language. i've been keeping an eye on it ever since, hoping it would go fully open, but it never happened.
Notice that those are tools which specialize in specific domains (e.g. linear algebra), not just a general purpose programming language. And those tools are even higher priced without the academic discount.
No one has ever built a significant business on a new general-purpose programming language. It doesn't seem to be due to the relative lack of IP protection on languages either. It may be that it's simply not a problem worth solving, there's no solution worth marrying a vendor over, or there are too many language designers.
Rebol has so many interesting ideas in it that it really could have carved a respectable niche in the world of networked programming languages. Just having sensible pragmatic built-in types for email and money was enough to bring it to my attention.
It's is a very odd but amazingly powerful little language worthy of study by anyone interested in language design. It has been a big influence on me even if I have never used it for anything 'real'.