fractalnavel

i use a dotted date notation most of the time these days (2005.12.03), and it occurred to me how messy these mixed base numbers are, especially when you consider time as well.  yyyy.mm.dd hh:mm:ss.hh - that's base10.base12.base30sortof base24:base60:base60.base10.  now, there's no particular reason why we can't use a fully base10 time of day (what people associate with “metric”), but the year.month.day relations are determined by the particular characteristics of our astronomical reality.  weeks are an interesting anomaly in themselves, an interesting historical accident of the arrangement of the local system combined with our evolving sensory (what we can see), social (when calendars became important) and neural processes.

the use of base10 in the first place is associated with the physical reality of our anatomy, or at least that's how the reasonably credible story goes.

ultimately, our preferences to preserve anachronisms is purely arbitrary.  there is no objective reason to maintain such (outside of considerations of the operation of a mechanistic universe, of course).  in an artificial environment disconnected from the chance arrangements of physics there's no reason we can't construct our temporal relations as we best see fit.

and the fit is determined by our evolutionary history, and so here we are.  perhaps someday - or somewhere - the fit will be different.

i, for one, think that using an astronomical year as the basis for many non-seasonal business dealings is pretty burdensome and downright annoying.  like filing taxes.  can't we disconnect this from the earth's orbit and the state of vegetation ?  why not file every five years ?  or every 100 days ? an objective analysis of economic optimization would likely result in something entirely different - although it might remain the same for farmers, etc.

and yet the results of objectivity must consider specific instantiations of subjective reality.  which kind of starts me thinking about an old project - prime free or other artificial numbering or metrics, designed with specific features only.  ideally, these would not be isomorphic to existing systems.  a disconnected category ?  and can unexpected features be prevented - ie, eliminate emergence ?

and what about numeric representations ?  a sequence of modular arithmetic applied to coefficients of polynomials in base x_i.  surely we can do better than that.  so much of our perceptions and conceptions are controlled by the distinctions we do - and do not - make.  what of a represenation-free mathematics ?