- David H. Ackley, Trent R. Small, Indefinitely Scalable Computing = Artificial Life Engineering
- Lance R. Williams, Self-Replicating Distributed Virtual Machines
presented at ALIFE14.
Both articles look great and the ideas are very close to my actual interests. Here is why:
- it is about distributed computing
- using actors
- some things which fall under functional programming
- and space.
The chemlambda and distributed GLC project also has a paper there: M. Buliga, L.H. Kauffman, Chemlambda, universality and self-multiplication see the better arXiv version because it has links inside: arXiv:1403.8046.
The resemblance with the mentioned papers and the chemlambda and distributed GLC is that our (fully theoretical, helas) model is also about distributed computing, using actors, lambda calculus and space.
The differences are many, though, and I hope that these might lead to interesting interactions.
Further I describe the main difference, with the understanding that all this is very new for me, a mathematician, so I might be wrong in my grasping of the MFM (please correct me if so).
In the MFM the actors are atoms in a passive (i.e. predefined) space. In the distributed GLC the actors have as states graphs called molecules (more precisely g-patterns).
[Here is the moment to thank, first, to Stephen P. King who noticed me about the two articles. Second, Stephen works on something which may be very similar to the MFM, as far as I understand, but I have to strongly stress that the distributed GLC does NOT use a botnet, nor the actors are nodes in a chemlambda graph!]
In distributed GLC the actors interact by message passing to others actors with a known ID. Such message passing provokes a change in the states of the actors which corrsponds to one of the graph rewrites (moves of chemlambda). As an effect the connectivities between the actors change (where connectivity between an actor :alice and :bob means that :alice has as state a g-pattern with one of the free ports decorated with :bob ID). Here the space is represented by these connectivities and it is not passive, but an effect of the computation.
In the future I shall use and cite, of course, this great research subject which was unknown to me. For example the article Lance R. Williams Robust Evaluation of Expressions by Distributed Virtual Machines already uses actors! What more I am not aware of? Please tell, thanks!