Observations

I just took a shower, and while I did, I was thinking about creativity. And bit decay.

I mentioned in an earlier post how I thought it would be most excellent to be able to see the possibilities inherent in data. I referenced the fellow who used data from multiple ships to conclude what the fastest transit path across the Pacific for sailing ships might be. I didn't mention the comment about how WalMart's data analysis showed that not only did they have higher than normal sales of flashlights and survival equipment, when a hurricane loomed -- that's to be expected -- but also Poptarts; reflection makes that a 'well, of course' observation, but the trick is doing it up front. Holmes might have scorned reasoning in advance of data; this is reasoning in advance of analysis. Is it possible to teach this ability? I want to say that I think it must be. To rely purely on the corp of deep thinkers seems -- well, not silly, perhaps, but wasteful. We don't expect that only the people at Frog Design can come up with elegant solutions to design problems (real world design problems, not art-for-the-sake-of-art design problems) any more than we say that only Warhol or Picasso can create pleasing images. Certainly, only that caliber of person can create capital-A Art, but anyone with a decent eye can snap a pleasing photo. If you train that person in techniques of photography, just the mechanics can improve the output. Don't shoot into the sun; take a breath and hold it before releasing the shutter. And more: the rule of thirds will lead to more pleasing photographs; knowing about Rembrandt lighting will result in better tonal quality. None of this is Capital-C Creativity; its more like 'enhancing the likelihood of creativity in daily activity'. Can this be trained? Can we suggest to people situations where they might find it of use, and a general idea of how to do it? I want to believe that it can.

Bit decay. It's a phrase that computer people use, and though I suspect that the deep thinkers there mean something different, for me it means the likelihood that a file which was perfect yesterday has something happen to it in the last twenty four hours. That something might be a stray electron here, a solar flare there -- but the ones and zeroes are no longer exactly as they were. In many, perhaps most cases it doesn't matter -- the image is essentially the same; the checksum corrects the values. But what about when the image isn't the same -- suddenly, a JPEG is unusable because it's shifted, somehow; the checksum couldn't tell me what the value should have been. What then? It's that unsureness that leads us to make redundant backups. And leads others to just not want to trust their data to computers. We've got an estate plan from our lawyer; he's given it to us on CD, and on paper. Guess which one we trust more to be usable in five, ten, thirty years?