Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Because They Have To ]   |  Signal Tasking  |   [ Downtrodden Rock Stars Unite! ] >>

Weekly Column

Signal Tasking: Steve Morton Proves That It Doesn't Take Thousands of Engineers or Billions of Dollars to Build a World Class Chip Company

Status: [CLOSED]
By Robert X. Cringely
bob@cringely.com

Have you noticed how the sports page has become the athletic business page? So many stories are not about the games at all, but about contracts, winnings, and bonuses. And it's not just in the U.S., but everywhere. We seem to be losing the love of the game, and replacing it with the love of the game biz. And yet the same thing happens in so many other industries. Just look at the Microsoft antitrust case, or even this week's maneuvering in the Recording Industry v. Napster civil case. We must love this infighting because we certainly read a lot about it.

In case you missed it, Napster, the MP3 sharing startup founded a moment ago by a 19 year-old, has a new lawyer — David Boise, the humbler of Microsoft. Boise, who is anything but conservative, has a new legal strategy claiming that since individual copying of music for a friend is widely accepted and not prosecuted, a web site that aggregates such behavior cannot be considered illegal either. If he can convince the Federal Court in San Francisco to go along with this idea, some folks think it might mean the end of copyright laws as we have known them.

Yeah, sure. If I drive over the speed limit and get away with it, does that mean speeding is no longer illegal? Or is it required that NOBODY gets a ticket before we can claim a law is no longer in effect? For some reason, this reminds me of the guy who parked his delivery trucks on the side of the parking lot owned by his neighbor, then successfully argued that by allowing him to do so, the neighbor had effectively abandoned the property to him. This is a true stor,y and is a good argument for being a bad neighbor. So will we see record companies suing teenage music pirates (OUR KIDS!!!) just to keep from losing their intellectual property rights? It could come to that. I like the idea of a kind of reverse class action lawsuit where the class of record companies sues the class of pirates. No, on second thought I don't like that. But as a guy who makes his living from intellectual property, I have to say that stealing is stealing no matter what dorm you live in. This doesn't mean that the record companies won't have to come to terms with the effect new technology is having on their ways of doing business, but they deserve to be paid for their property.

So let's get back on a positive note with my continuing series on exceptional startups. This week, it is Oxford Microdevices, the biggest chip company having its world headquarters in Monroe, Connecticut. Oxford, with fewer than 10 employees, is also the smallest chip company having its world headquarters in Monroe, Connecticut. Certainly, Oxford teaches us that you don't have to be a big company to be a good company, and you don't have to come to Silicon Valley to do good work.

Oxford is run by Steve Morton, and seems to be mainly staffed by folks named Morton. Steve is a chip designer who specializes in Digital Signal Processors (DSPs), those specialized chips that are used to power cellphones and graphics adapters and essentially make digital multimedia possible. If it involves signal compression, decompression, or manipulation, a DSP is involved. By far the biggest company in the DSP business is Texas Instruments. Philips is also a big player in that business. Oxford Microdevices is probably the smallest.

But in this case, small is good. Just as Be was able to build a world class operating system (the BeOS) with fewer than half a dozen programmers, Oxford has been able to design world class DSPs on a very low budget. This is the advantage of having no legacy customers, applications, or products with which to maintain compatibility. Life is so much less complex when you start with a clean sheet of paper.

Oxford's A236 DSP is shipping now, and its A436 chip will be ready this summer. These chips will shortly begin appearing in a variety of products, but much of the early product development has been in the area of fingerprint recognition. There are handguns that use Oxford chips, and a sensor in the trigger to ensure that the gun will only fire for its owner. Fingerprint recognition that takes only as long as it takes pull a trigger is a remarkable technical feat. Doing it for less than $50 is even more amazing. Alas, what would be most amazing of all would be having this capability actually appear in guns that are in mass production. Despite years of work and lots of positive publicity, Oxford has been unable to interest a single gun maker in adding this capability. They won't do it until they are made to do it.

This field of fingerprint recognition is fascinating and sobering. There is so much press coverage of children being kidnapped and so many efforts to protect our children through techniques like fingerprinting that you'd think computers would play a role. Not so. An interesting fact about fingerprints is they don't add features as we grow. Rather, the fingerprint of a small child has all the same patterns as the fingerprint of that child as an adult. The kiddie print is just scaled down. But this has a negative effect on fingerprint recognition. You see, the sensors used in computerized fingerprint systems have a resolution of 500 dots-per-inch. This is plenty good for identifying adult fingerprints. But try to look at the finger of a five year-old with these systems, and it simply doesn't work. The ridges in a child's finger are too close together to be resolved by a 500 dpi sensor.

What fingerprint systems for schools should have is 1000 dpi resolution, but that's a much more expensive sensor. And since it produces four times as much data from each scanned print, it requires a more powerful processor. This is no problem for Oxford, which has some of the most powerful cheap chips in the world, but then the world probably doesn't know about Oxford. And that's the problem.

Working with a core of third party developers, the Oxford chips are being built into many remarkable products. What makes these products remarkable is that they are able to take advantage of cheap but incredibly powerful processors that make it possible to save money through additional processing steps. Just like the Hubble space telescope uses signal processing to overcome its optical problems, Oxford chips are being used in digital cameras to run not only the whole camera, but to pre-process the image coming through a cheap plastic lens, making look as if it had come through an expensive glass lens.

These chips, which can be programmed in C, are powerful enough for real-time MPEG-4 encoding at full screen rates. If we ever have TV-quality videophones, they will use chips like these. And we might see these chips inside TVs themselves. Oxford's A436 chip is comparable in performance to the new MAP-CA DTV chipsets from Equator Technologies, except Oxford's chip produces one-fifth the heat, uses one-fifth the power, and has one-fifth the cost. Don't expect to see Equator's chip in a battery-powered TV anytime soon. Oxford's chip could do it tomorrow.

My point here is that individuals can still create world class technologies. Oxford has no venture capital. And the advantage of being a little guy is flexibility. If some startup had raised $30 million to build smart gun technology, the realities of the firearms industry would have put them out of business for lack of customers. For Oxford, the pig-headedness of gun manufacturers just means turning to some other opportunity like smart cameras or teleconferencing. When and if the gun industry ever comes around, Oxford will still be there, ready to supply smart gun technology.

I like little companies like this, and the clouds of independent product and application developers that form around them. It is from outfits like Oxford and its developers that some of the important technologies of the future will arise. It doesn't take thousand of engineers and billions of dollars to do good work. Clever design can still substitute for brute force. In a world obsessed with gigahertz clock rates, for example, Oxford's A436 chips runs at 66 MHz, making it useful for battery-powered applications.

The startup in a garage is far from dead.

Comments from the Tribe

Status: [CLOSED] read all comments (0)