Photonic Breakthrough Allows 'Lab-on-a-Chip'
Roland Piquepaille writes "Georgia Tech researchers have shrunk an optical device called wavelength demultiplier (WD) by combining into one crystal three unique properties of photonics crystals. This optical discovery opens the way to sophisticated and cheap bio-sensors mounted on 'lab-on-a-chip' devices -- sensors to run blood tests, detect chemicals in water supplies or for drug testing. Their new WD is less than a millimeter in all dimensions rather than the several centimeters of other currently available WDs. And it should not cost more to produce."
Now the key components for processing light have been miniturised, it should be possible to use this for the basis of a simple optical computer.
From the sources below:
http://www.theregister.co.uk/2006/08/03/photonic_
So... it seems like the photonics are ready. We just need engineers to step up to the challenge.
Signature v3.0, now with 42% less memory usage.
drug testing
The War On Drugs is sooo last century - didn't they get the memo that we're fighting the War On Terror now? They haven't sensationalised it enough - they've got chemical attacks on water supplies in there, but it's too subtle - where's expolsives detection at customs on their list?
As a scientist, this is what I don't really like about scientific journalism. Like the 'New breakthrough in fighting cancer' titles, etc. etc. These are laboratory research developments and will take at least 10 years to evaluate, some of them will end up being impractical before ever being put to use.
I think that scientific journalism should be more than just a PR machine for research labs. Of course they want the message out that they're doing nice stuff, but as it are all just small pieces of advancement, don't bring it as if you just developed a working nuclear fusion reactor ready to connect to the powergrid. Show that you're doing someting nice, what it can do, what the scientific/technical genial idea is that was done to get it, and in what frame we should see it, that should be enough.
molmod.com - computing tips from a molecular modeling