Is the contest constrained to the Turing machine model?
Not at all. Actually we encourage people to follow other models of computation (sk combinators, rewriting systems, cellular automata, tag systems, neural networks, register machines, etc.) and write down the shortest implementation you can come up with and write an encoder/decoder to accept the required (Turing machine-style) input syntax and fulfill the required output format of the contest. Of course that may undermine any possible length reduction ensuing from the chosen model, but that is precisely the issue we want to address. Isn’t it ironic that the function to implement was only primitive recursive? A: Yes it was. Implementations before January the 20th. 2009 are loop programs that always halt. They are only universal when the step limit, that was required as input by the former rules, is dropped. To do so, in terms of coding is quite easy since one would just delete the aditional condition for reaching the given maximum of steps. The possibility of unbounded computation is ess
