GeoHashing represents an interesting use of real-world oracle-based computation. In short, DOW opening index on a given day is used as the oracle input sequence to compute a Lon-Lat coordinates for a meeting place for ad-hoc assembled adventure seeking participants. The meeting locations cannot be computed ahead of time, and Friday DOW opening index can be used to generate meeting locations for the weekend days. The initial idea is attributed to xkcd webcomic which gained popularity in certain circles. To get a taste you can check for e.g. On purity.
In the context of theory of computation, geo-hashing represents an interesting real-life computation process that cannot be modeled by a Turing machine (unless someone turns out with a Turing machine capable of modelling the stock market and predicting exact DOW indices). The only existing theory of computation applicable here would be the oracle-based machines, or O-machines in Turing nomenclature. Theoretically, given appropriate oracle, an O-machine is capable of computing “more” than any normal Turing machine can – which is simply demonstrated by the ad-hoc meetings through geo-hashing. Can such o-machine-based computing schemes will become abundant in Web 2.0 and post Web 2.0 era? What other useful things could non-turing-computable real-life o-machines be used for? The linkages between computing and real-world are starting to be washed away – computing is becoming part of real-physical world, tightly coupled with real physical events and phenomena. O-machine-based computing might become the dominant computing paradigm, and things will get really interesting then – because most if not all limitations of Turing-machines will cease to be applicable anymore. Will this put Fredkin’s Turing-machine equivalent universe into question?