The only thing we can be sure of about the future is that it will be absolutely fantastic. So if what I say now seems to you to be very reasonable, then I have failed completely. Only if what I tell you appears absolutely unbelievable, have we any chance of visualizing the future as it really will happen.

- Arthur C. Clarke (h/t Brin)

Saturday, December 1, 2007

Words I'm Giving Up for the New Year

<seinfeld>

I'm going to say a word. You tell me the first thing that comes to mind. Ready?

Model.

What's that mean, "model"?

Did you hear the one about the software engineer, the climatologist, the psychologist, the statistician and the fashion photographer?

I am swearing off that word. It's got so many meanings. Any sentence with the word 'model' in it is likely to get you into trouble.

I'm giving the words "uncertainty" and "feedback" a very dubious eye as well. I am currently inclined to think they are also not usable in mixed company, but I'm not sure I can manage to go cold turkey on those. No, you shouldn't be too hard on yourself at New Year's.

Oh, and "error bars". What's up with that?

I never want to see those error bounds things again without a clear description of what you mean by it. "This is the manufacturer's spec on the measurement instrument" would be fine, but it's been a long time since I've seen one of those.

Does your error bar mean "that's how confused I think I am"?

I mean, seriously... Are you 100% sure that you're 75% sure of that? Rilly? What is that, an "imprimatur"? Do you guys get them from your spiritual leader or what? Because, you know, we don't get those.

Seriously, if I had some error bars I'd be slapping em on everything in sight; if you know how to detect a sure thing don't you think you should be out there making the most of it?
</seinfeld>

If science worked better, there would be a definition of the word "model" like that of the word "energy" which means something very specific in formal conversation, but there isn't such a definition. The word is a complete mess of vaguely associated concepts. I'm serious that dropping the word "model" from my vocabulary has greatly clarified my writing.

To climatologists looking askance at me here, the word you are probably casting about for is 'simulation'.

2 comments:

John Fleck said...

Michael -

The 8 November Nature had a review of a book that might interest you in this regard. The book's entitled "Science Without Laws", Duke University Press, edited by Angela Creager. The review made it sound like a fruitful exploration of the various ways "models" (loosely defined) are used in a variety of scientific settings.

EliRabett said...

Not simulation, because models can exist prior to the data and independent of data.