Jon Harrop skrev:

...

I was unable to find the quoted phrase via google, so I assume
that the quotation marks were not to be taken literally? (:

Surely that Erlanger was mistaken, but I hope that this is
not a common misconception (it's certainly not a commonly
held opinion on the erlang mailing list).
If it is, let's clear things up.

I wrote the following in a fairly well cited blog article
on erlang-style concurrency:

"There will always be limits. Erlang was designed for agent-style
concurrency; not for massive data parallelism."
(http://ulf.wiger.net/weblog/2008/02/06/what-is-erlang-style-concurrency/)

Also, those who have read up on erlang's history should also
know that experiments with Parlog were part of the early
evolution.

"1985-86
Experiments with Lisp,Prolog, Parlog etc. Conclusion: [...]
It must also have a granularity of concurrency such that one
asyncronous telephony process is represented by one process in
the language. (Rules out Parlog.)"
(http://www.erlang.org/course/history.html)

Finally, we have Joe Armstrong's History of Erlang (page 7):

"We started looking at languages like Parlog, KL/1 and Strand
for possible inspiration.
This actually turned out to be a mistake. The problem here has
to do with the nature of the concurrency. In the concurrent
logic programming languages, concurrency is implicit and
extremely fine-grained. By comparison Erlang has explicit
concurrency (via processes) and the processes are
coarse-grained."
(http://portal.acm.org/citation.cfm?id=1238844.1238850)

BR,
Ulf W