jbn@glacier.STANFORD.EDU (John B. Nagle) (04/08/89)
Proper science has recourse to experiment to decide issues. This
makes it possible to establish a base of facts firmly grounded in
repeatable experiment upon which one can build. This is one of the
main reasons that physics moves steadily forward.
AI needs a sound experimental basis. Within five to ten years, it
will be given one. More on this when comp.ai.digest comes back on line.
John Naglegilbert@cs.glasgow.ac.uk (Gilbert Cockton) (04/10/89)
In article <18261@glacier.STANFORD.EDU> jbn@glacier.UUCP (John B. Nagle) writes: > > Proper science has recourse to experiment to decide issues > AI needs a sound experimental basis. > > Before rushing in, perhaps we could discuss the nature of acceptable and convincing experiments in cognitive science. They are not easy to design, are rarely designed, and involve more than making something and then shouting that it's doing something. -- Gilbert Cockton, Department of Computing Science, The University, Glasgow gilbert@uk.ac.glasgow.cs <europe>!ukc!glasgow!gilbert