Featured Post

Pinned Post, A Policy Note:

I have made a decision to keep this blog virus free from this point forward, at least until the smoke clears. This is not a judgement about ...

Sunday, February 22, 2015

Local Maxima

In mathematics you find this idea of a "local maximum" which is, roughly, a place where everything nearby is lower down. Some of us call this a "hill".

The point is, though, that it's the highest place nearby, but not necessarily the highest place there is, or even the highest place in a somewhat larger area.

The thing about local maxima is that these are what iterative processes of improvement find. If you're out walking and you try always to walk uphill, you'll find a local maximum. Sometimes you have to go downhill and cross a valley to find a higher peak.

So you just got a camera and you're taking pictures. Probably they're not that great. You might start out making sweeping changes every day but pretty quickly you're likely to start making small changes. You work on getting focus right. You're engaged in an iterative process of improvement. You change a little thing, you see if the result is better. If yes, keep going. If not, make a different, small, change.

If you're in a camera club or online sharing venue, or really anything that lets other people who don't necessarily love you tell you stuff about your pictures, they're going to help. They'll offer single suggestions "too dark" and so on. They will create an environment of what "uphill" means for you, and they'll guide you iteratively, step by step, up that hill.

So you're going to find a local maximum. At this point making changes will tend to make your pictures worse (in some sense).

Let us imagine the world of all possible photographs. There are higher areas that represent "better" pictures, and lower areas that represent "worse" pictures. It doesn't matter what you mean by better or worse. Whatever you like, really. Popular? Artistic? Thought provoking? Anything. It doesn't matter.

If you find yourself on a hill labelled "flickr Explored" you're at a local maximum. You are simply not going to iterate your way, tentative step by tentative step, to something better.

If you want to make powerful art, or whatever, anything that is both "good" by any definition you like, and which is also not instantly recognizable as "flickr Explored", you've got two choices:

  • Iterate your way step by step through some valley of bad pictures, trying, probably at random, to find a new, higher, peak somewhere on the other side. Ugh.
  • Line up on a higher peak, back off 10 steps, run like hell, and take a gigantic flying leap.

Either way, you have to make a radical change. Small changes are just going to produce crap.

Don't like where you are? Stop fiddling around, and start lining up on some higher peaks.

And get ready to fall.

No comments:

Post a Comment