Caltech creates 1um resolution microscope cheap enough for home use.

Yes, cheap enough for home use, but not necessarily suitable for home use.

The idea is that instead of using lenses to get a magnified image of something at a distance, you arrange for your sample to go really close to a digital-camera-style sensor without any lenses. You get multiple images by sending the sample past several bits of your sensor. Then you do some computation (the article doesn't say what for, but I guess the idea is to increase the effective resolution by having those multiple images, and maybe also to remove diffraction artefacts).

All very neat, but only suitable for capturing images of things you can actually do this to. The idea of the system described in the article is that you can do it to cells suspended in fluid (water? something else? I don't know) if you're clever about controlling how they move over your sensor.

But if what you want is a greatly magnified image of a fly's wing or a bit of paper or an integrated circuit, or (I'm guessing here) even something that is (as per the design) suspended in fluid but that might have things larger than cells that could block the channels through which the fluid moves, then you're out of luck.

I'm sure it can be generalized somewhat, but there are some limitations that seem really hard to overcome. You absolutely have to get the sample very close to the sensor. You absolutely have to illuminate it transmissively rather than reflectively. (At least, that's how it seems to me.)

In this New Yorker article: http://www.newyorker.com/reporting/2008/05/12/080512fa_fact_... , there's a discussion about a novel cancer therapy for which this microscope is well-suited. The idea is that, when cancer metastasizes, that each of the cancer cells circulate in your bloodstream thousands or tens of thousands of times, so you've got plenty of time to nix them, if you can find them. If you could design a device that would shunt some of the blood out of and then back into a blood vessel, check for cancer cells in between (if you could develop a recognizer, which is apparently feasible), and remove the cells one at a time, you might be able to slow the disease long enough for other treatments to run their course.

Something's wrong with this story's comments here. What's going on?

Links to real info:

http://www.biophot.caltech.edu/research/ofm/ofm.html

[edit, more better:]

http://www.rsc.org/delivery/_ArticleLinking/DisplayHTMLArtic...

Can it fill a house with popcorn?

i heard iDealsChina already has leaked pics of an iphone with this embedded

Wow, I have material for at least six months' worth of small penis jokes now.