Comment Re:Fundamentals of Comp Sci (Score 1) 315
^^^ Actually, the attempt was to simplify something in each version. Oberon 2013 is perhaps the simplest one yet, but there's really two lineages to watch, and most people would probably find the BlackBox Component Builder much more palatable. (Although it's in a need of a massive overhaul, if you ask me - I'm actually considering looking into trying to graft the match-bounded polymorphism approach onto it, as outlined in Kim B. Bruce's Foundations of Object-Oriented Languages, because it just might turn out to be a good fit for the language. It's a massive effort, though; a whole 3.5% of the effort needed to write a correct C++ compiler
You know, the "burn the diskpacks" approach was actually what got Smalltalk in the 1970s in a short period of time into the forefront of software systems. Outside of Lisp machines, I'm not aware of anything that advanced in that time period. But after they set the language and system design in stone to commercialize it, most of the advancement has stopped, and new aggressive designs started later catching up (somewhat, at least). The same thing has happened with the (post?)modern computing systems like Linux or Windows that succeeded it in many of its domains: once you accept the status quo, the only way to "improve" the system will involve piling a complex layer on top of the existing system to "fix" the deficiencies of the layer beneath it at least for new applications. Significant new progress in the field suddenly becomes almost impossible under these circumstances.