Thursday, 23 August 2012

in praise of waste paper

I have a stack of "rough paper": paper printed on one side but no longer needed, that I keep by my desk for scribbled calculations and diagrams, for when I am thinking.  I've never really considered this, except occasionally to think that I should actually scribble these notes into my daybook as I go, so that they are part of my technical log.  That has always felt wrong to me for some reason.  Now I know why: Venkatesh Rao has the same experience, but understands it:
I can’t really work with ruled paper or blank two-sided paper. Both are pristine resources that I feel guilty about wasting with my manic scribbling of mostly useless thoughts. ...
... the reason I work best with one-sided printed paper is that it has already been bad-wasted. It’s going to the recycling bin anyway. So it is a particularly liberating medium to work with. More than liberating. You can feel virtuous because you are effectively redeeming bad waste.
Yes, exactly this.  I can draw a small diagram in the middle of an A4 sheet, dislike it, discard it, and draw again, and again, with no feelings of guilt (like Venkat, I use other people's discarded paper).  This freedom to redo things is very liberating.  If I was drawing in my daybook, or using some pristine pad of blank paper, I would feel constrained to be more frugal in my explorations.  As it is, I give the paper no thought whatsoever (except maybe a small glow of productivity when I discard a great wodge of it in the recycling bin at the end of the day: look at all that work I must have done!).

Venkat's post is about waste-enabled creativity: "civilizations are defined by the resources they can waste".  He talks about exploration-exploitation tradeoffs, where being able to waste resources allows cheaper, and hence more, trial-and-error exploration, and hence the discovery of potentially better solutions to exploit.

It's not just civilisations.  Biological life can be similarly defined.  Computer Scientists often (incorrectly) assume that biological evolution is a great optimisation algorithm.   This is incorrect for two reasons: (1) it's not "great"; it's wasteful: millions are born to give a varied pool from which a few fit are selected; (2) it's not an optimiser.  Biological life "wastes" enormous quantities of resource (a million cod are spawned so that two may survive to reproduce); it explores the vast evolutionary landscape like mad, and is fantastically creative as a result.

We Computer Scientists love bio-inspired algorithms, because of the (assumed) robustness of biological organisms.  But if only, we often wish, especially when we are being "good" frugal engineers, if only we could have all those lovely biological properties without all that waste.  Hmm.  Maybe in our attempt at software engineering and Artificial Life, the problem is that we are not wasting enough, and that we are wasting the wrong things

No comments:

Post a Comment