Measure what can be measured, and make measurable what cannot be measured.
A modification of this famous quote was the title of an interesting paper I read on compressed sensing a couple of days ago, “Measure what should be measured”. In today’s world of data explosion it is a curious statement. The real thing that might apply is “measure everything, look at what is important”. This might have real application to control the expansion of data. Compressed sensing might have a great deal to say about how to do this.
I started thinking about computing. What do we do today, do we simply “compute what can be computed”? Shouldn’t we be “computing what should be computed”? or better yet compute what is important. Like data, we perhaps will “compute everything and look at what is important”. Again, the philosophy and methodology of compressed sensing might apply to getting there.
One route is efficient, but extremely difficult, while the other is wasteful and potentially tractable. Something needs to happen, the future of computing depends on the answer.
The Regularized Singularity 