OK. .... judging by the downvotes, I've gotten some technicality wrong. Anybody care to tell me what the problem is? Or did tunable lasers kill someone's brother?

Tunable lasers are expected to be a big business for telecommunications equipment, for sending signals down fiber optic cables ... unfortunately this application uses infrared wavelengths, so you won't be able to watch any colors change, which takes a lot of the coolness out of these things.

It's assumed that telecom lasers will all be tunable someday, and designers will simply have the option of using the tunability or not, depending on whether they need it. So the dream is that "tunable laser" will be a redundancy, at least in telecom circles.

Tunable lasers would be useful here because of wave division multiplexing (WDM), a technology that lets multiple wavelengths of light share a fiber-optic cable. The plus side is that this has increased the supply of bandwidth tremendously, particularly in the core of the Internet, without having to install more fibers. Unfortunately, you need a laser for each different wavelength -- and then you’ve got to keep spares of all those lasers. It adds up; as of 2001, WDM is up to 160 wavelengths per fiber IIRC.

Tunable lasers can transmit a certain number of wavelengths in a predetermined spectrum. Now you need only one laser to send your WDM'ed information, and you only need to buy one type of laser to keep in the spares drawer.

Drawback: These tunable lasers are new -- I don't think they started shipping before 2000 -- and therefore pricey. And telecom companies aren't willing to pay that much to get rid of this inventory/spares problem. I mean, let's face it: Tracking 160 parts is annoying, but it's not rocket science.

Price will come down over time, yes, but there's another problem: There's no standard way to make a laser tunable. During the business boom of 1998-1999, dozens of little startups got funded to build tunable lasers. The result is that at least four or five different types of tunable lasers exist, and they aren't all the same shape and size. So you have to build your board for a particular tunable laser and hope that that company doesn’t die on you. Luckily some tunable-laser companies are starting to agree on package sizes and shapes.

Other drawbacks: In some cases, tunable lasers have lower output power than regular lasers, making it tougher to send signals for long distances. This matters to folks like Qwest or British Telecom trying to build continent-sized networks, because optical amplification is expensive. Also, some tunable lasers use moving parts. If you're familiar with MEMS, you’ll know that these moving parts are pretty reliable, but some people can’t shake the feeling that they might wear down over decades of use.

There are very small tunable lasers that are very reliable and behave basically like a regular laser. But they can only tune over a range of 4 or 16 nanometers.

Companies making this stuff (with no guarantees as to how successful or competent they are!): Agility Communications // Bandwidth9 // New Focus // Coretek (now owned by Nortel) // Iolon // Altitun (owned by ADC // Tons of others.