MIT AI Lab Memo No. 750 -- November, 1983
"Design Issues in Parallel Architectures for Artificial Intelligence"
by Carl Hewitt and Henry Lieberman
http://citeseer.ist.psu.edu/530318.htmlThis short (11 pages) paper describes the differences between traditional
computer languages and those based on the Actor model, making a case for the
application of Actor-based systems in AI. Their description of "open
systems", characterized by "continuous change and evolution", "absence of
bottlenecks", "decentralized decision-making", "arms-length relationships"
(aka: no shared state), "perpetual inconsistency among knowledge bases" and
"need for negotiation among system components", matches closely the kinds of
challenges we face in todays distributed multi-processing environments.
Hopefully this paper will assist in clarifying the relevance of the
Actor-model to addressing these challenges.
Enjoy!
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://redfish.com/pipermail/friam_redfish.com/attachments/20080325/21544bd5/attachment.html