Erlang is a very nice language, but nevertheless it does not include any feature that can be considered as an innovation over older languages.
A programming language does not need to include anything new to be a better language, it is enough to offer a combination of good features that do not exist together in competing languages.
"Parallel threads of execution that explicitly listen for and send messages" as the authors says, already existed in Ada (1979).
Moreover, such concurrent programming already existed in PL/I, in 1965. PL/I, in 1965, was actually much more convenient than the supposedly modern POSIX pthreads, because it was the first to include the equivalent of WaitForMultipleEvents, which is sorely lacking in POSIX.
Interesting points about PL/I. The more I hear about the 60s/70s in computer science, the more I wonder what exactly happened. I've heard a lot about interesting things existing at this point, and people just rediscovering them recently. I don't know much about this time, and know even less about other engineering disciplines, so I'm not sure if it's specific to computer science. Or maybe open source means that we actually hear about it, and when 10 companies design the almost same kind of wheel in a NIH here fashion we don't hear about it? I wish we had a bit more of a focus on history in our field.
What happened was that some programming languages, like PL/I, were available only on very expensive computers, e.g. IBM mainframes.
The languages available on cheap computers, e.g. BASIC, Pascal or C, had less features, especially features for parallel processing, which were not useful on cheap hardware.
In time, the cheap computers became more powerful than the old supercomputers. Then many of the features formerly available only on powerful computers were added to the new popular programming languages, but not all of them.
While what were considered large languages during the sixties, e.g. PL/I and ALGOL 68, had a few serious flaws, they also had many nice features that are still not present in the most popular programming languages of today, mostly as a consequence of the fact that while C has taken most of the features it added over BCPL and B from either PL/I or ALGOL 68, it simplified those features a lot or even crippled them compared to the original, in order to allow implementation on much cheaper computers.
Later languages attempted to be better than Pascal or C, which was a very easy target, but their designers did not study what was available in earlier languages, to be also better than that.
A programming language does not need to include anything new to be a better language, it is enough to offer a combination of good features that do not exist together in competing languages.
"Parallel threads of execution that explicitly listen for and send messages" as the authors says, already existed in Ada (1979).
Moreover, such concurrent programming already existed in PL/I, in 1965. PL/I, in 1965, was actually much more convenient than the supposedly modern POSIX pthreads, because it was the first to include the equivalent of WaitForMultipleEvents, which is sorely lacking in POSIX.