Slashdot Mirror
A House Divided: UWB's Double Standards
Mai writes "What happens when two coalitions within a standard come into conflict, and it doesn't get resolved quickly? The ultrawideband technology standard shows you."
← Back to Stories (view on slashdot.org)
Standards in Artificial Intelligence are just as difficult to set. No particular body or coalition has the God-given authority to set standards, so we just have to let the standards evolve.
That's what you get when you put the standard before the technology. Step one: let companies use whatever standard is convenient for them, and sell their products to whomever they can convince to buy. Step two: once the market has tested the products, standardize based on best current practices.
Sure, it has the net result of lots of poor guys owning a collection of relatively useless Betamax videos, but really, I'd rather own an obsolete product because it made its best shot and failed than own a mediocre product because it conformed to a political compromise that had no market time behind it. (And furthermore, it encourages the Betamax owners to switch to DVD more quickly than the content and universally supported VHS owners, thereby even further spurring development.)
For real life examples, read some W3C Recommendations. The ones that were presented as ready-made standards before any market was actually implementing them (like PICS) are lovely pieces of technological poetry. The ones which were widely tested in the market and implemented first, even if in lots of non-compatible subversions, and only then standardized (like HTML), on the other hand, are actually used.
What he wants is more important that what I want. What he wants is also more important that what you want.
There's a book titled "In the Blink of an Eye," which details the Cambrian ecological explosion, which was precipitated, by (you guessed it), the emergence of vision among vertebrates. I expect a similar explosion for machines. In the modern era, we have rather well advanced robots, in terms of data sensing and actuators, but really crappy AI and control. The availability of decent remote control requires wireless video (who want's a robot attached to the wall?). UWB (or perhaps one of its successors) will make robotitcs fessible. With the coming shrinking of the labor pool due to the baby boomers retirement, there will be a demand for labor-saving technology, a void which robotics will be poised to fill.
As a former employee at a UWB startup, I can say that the whole process is screwed up. You have companies that recruit uncles and aunts of employees who aren't even remotely technical and fly them around to IEEE meetings so they can vote for the companies chosen standard. You also have companies like TI and Intel with huge 802.11 patent portfolios who are interested in pushing the MBOA spec so that UWB basically dies a horrible death. Or at least that was the perception at our company.
In any case, this stuff is at least 2 years away from consumers(at least for a full 802.15.3a implementation) and will probably be made pointless with advances in other technologies. Then again I'm a bit jaded, having been through the start-up spin cycle and all.
If you dig far enough back in the history of this, there were quite a few different approaches. About four years back, there were four main groups. iNTEL figured they had to lay their stake early, so they tried to muscle the other three out. eXtreme Spectrum had almost enough horsepower in their design to survive iNTEL's tactics, but finally had to turn to Motorola to bail them out.
Why Motorola? Good question. Many of the core group at eXtreme Spectrum had just been laid off from Motorola, or had jumped before the layoffs.
Why does iNTEL want this?
This short-distance UWB is potentially going to replace all the wires on your information devices except for the power wires. If you were into patents, wouldn't you like a patent on wire? ("Patent on wire" was something of a buzz phrase during the infighting.)
(Personally, I'm wondering if the XS/Freescale technique techniqe might be beneficial in power wires, but I don't know what I'm talking about.)
If you ask me, though, iNTEL's idea of jumping from spectrum to spectrum seems to have the larger footprint, the higher susceptibility to eavesdripping and being dripped on, and the greater power requirements. It might scale to greater distance, but that might not be desirable for short-distance, high-bandwidth wireless.
The XS/Freescale approach of basically spitting raw bits into the air at pseudo-random frequencies and very low power should be familiar with a crowd that understands crypto. You remember the story about the actress and the piano player and an early patent in the field.
But, again, I don't know what I'm talking about, so my biases might be showing.