27 April 2007

When is information not information? (2)

Waking this morning, I find that Jim Putnam has, in the intervening eight hours, already responded to last night's post - which had taken me a laggardly three weeks to put up. I am suitably chastened...

Jim's systems analysis view of information is unarguable. It also illustrates the slipperiness of this whole topic.

A stack of printouts in the corner are no use to anyone, and therefore not information. The same, then, must presumably be true of a stack of sensory data in the corner of my brain which I cannot interpret - perhaps a set of unidentified sounds?

His comment about the distinction between hearing and processing (by extension, between any sensory input and processing) is right on the button. Almost everything we think we "see" is actually an internal result of processing.

The highly engineered Pentax lens on the front of my SLR is capable of resolving 12500 image points per square millimetre of film or digital sensor surface. The lens of my eye can only muster 64 at the retina; to make things worse, the image is focused through aqueous and vitreous humors, not to mention those fatty ropelike floaters - and most of the retina surface can't make full use of it anyway. And yet ... the image emerging from my SLR is to be measured (and usually found wanting) against the highly detailed and information rich image in my mind, not the other way around.

This is, of course, because my eye continually shifts to multisample the scene before it, and miracles of high speed image enhancement transparently assemble and deliver a real time processed result to me instead of the raw data. What I think I "see" is actually a sophisticated, software mediated, model.

All of which supports Jim's view: the image formed by the lens of my eye is no use to anyone. Only the processed model is useful information. As Dr C flagged up in his Information V, the processing starts immediately: the eye doesn't just passively pass on raw data, but processes it at a low level first. The uncertainty lies in when, exactly, the one (raw data) becomes the other (processed model)? I don't have an answer - I just ask the question, then walk away leaving somebody else to deal with it. Like Jim, I am learning from the discussion.

No comments: