Anyone who has done any programming that has elements of persistency (not just adding two numbers, but recording when that addition was performed) has experienced the challenge of time. It is a problem that exists on any scale, from human relevant scales (which online shopping order was placed first) to machine-relevant scales (which network packet was sent first). It is a problem that escalates as computer systems get larger (scale up) and faster.
This is an intruiging talk by Paul Borrill about the notion of time how it reflects on how we do computing (and how what we do is still shaped by the linear tape that was used when computers were born).It suggests that our linear approach to time is unfounded and causes much of the complexity we have to deal with in computing. Human beings are required to deal with this complexity. The result being that scalability is limited by what human beings can oversee / manager / administer. What if it was possible to create software (and hardware) in a different way. What if computing could be created without a “God View” – a sense of overseeing centrality, without monolithic source-of-record storage?
via Matthew Schutte and Arthur Brock – Ceptr / MetaCurrency