The Pebble and the Avalanche

The Low Emission Car Problem

Honda will release an automobile that uses hydrogen as fuel; the car will be available in Japan and in a few parts of California starting in the autumn of 2008.

Automobiles are part of a system that includes roads, traffic signals, repair shops, spare parts manufacturers, and — above all — fuel supplies. The world has a fine network of gas stations, and it's easy enough to run an electrical cable to an outlet to re-charge an electric car; but introducing an entirely new type of fuel means building entirely new and expensive infrastructure.

The creation of this new infrastructure is a problem in disaggregation that has yet to be solved. I suspect that the solution will lie in some interesting new thinking about what it means to be an automobile manufacturer, what it means to be a service station, and what it means to be a public utility.

Topics:  · government · technology

Link to this story · Leave comment or trackback

Twitter as a Technical Resource

If you haven't used it, Twitter is one of those "I don't think I get it" technologies. Leaving aside the cute terminology, Twitter provides microblogging. Ordinary blogs are short essays, such as this one; when I write a blog post it usually takes considerable time and effort. Twitter accepts only very short blog entries, 140 characters or less. These updates are sent to anyone who subscribes to them and can also be tossed into a large public pool for anyone to read. While this might sound sort of useless and derivative, it's anything but; I ask you to recall that blogging is really just a quick-and-dirty way to update web pages, and look what it has become now. Twitter lowers the barrier even further.

Is Twitter popular? Yes, extremely so; new media technologist Dan York has a Twitter account followed by over one thousand people. His "tweets" cover everything from his his latest technology thoughts to the local weather. (My account is the much same way.)

I've recently discovered that Twitter can provide an amazing technical resource: instant expert help from people who you didn't know even existed. Some individuals monitor the entire Twitter stream for certain keywords — and they might respond to your comments with extremely welcome help. Earlier today I noted in passing that I couldn't find a particular software function in the Ruby programming language; a few moments later Ivor Paul responded with a few well-chosen links to Ruby documentation that will cut hours off of my learning curve. And earlier this week Neil Edwards, twittering from London, gave me a leg up on finding the right software to make Ruby on Rails more useful.

I'm fascinated by the capabilities of Twitter. Twitter is disaggregated: Twitter allows access to Twitter as a a building block for other services. Now I've begun to wonder just where all this will lead.

Topics:  · internet · technology

Link to this story · Leave comment or trackback

Comments: 1, Trackbacks: 0

No Cure Yet for Software Rot

Every year at about this time I do a bit of community service: I have my computer system make outbound phone calls to friends and relations who want a reminder of one of their religious obligations that's rather easy to overlook.

This year I have a very, very severe case of software rot. "Software rot" isn't a term in any dictionary; it's a phrase that I made up but that any technical person immediately understands. If you put a piece of software aside and don't look at it for six months, you often have trouble starting it up again. The software rots, a problem just as real as if though you'd put away your saddle away after riding without cleaning it properly. In my case, the outbound calls seem to work, but — ironically — the quality-assurance subsystem that I put into place verify that the calls are actually successful is itself failing.

I mean this post as a cautionary tale for managers. This particular application, as are many of the applications I work on, is built of many different subsystems. Some pieces are written in one programming language (Python) and others are written in more specialized languages (CCMXL and VoiceXML). The software relies on a wide variety of Internet protocols to work correctly (DNS and HTTP in particular). I rely on components from outside vendors and I've installed new hardware that uses some apparently very touchy new software.

Although it would be ridiculously expensive to build all this infrastructure from scratch intead of using these coponents, using the disaggregated components does impose some other costs. Any significant project I build uses change control to keep track of the software that I create, but — unlike a commercial service — I haven't instituted controls on this project to track software modules from elsewhere and the various bits of hardware. And since one of the items in the mix is a Windows-based PC, I suspect that I know where the problem really lies.

My point here is that disaggregation brings both incredible costs savings and a certain responsibility. I will get this service up and running correctly again (hopefully sometime today), but if this were a commercial service I'd be in trouble. Just because a service worked six months ago does not mean it will work today. No software exists in a vaccum. Software-based services require constant maintenance to avoid software rot.

Topics:  · technology

Link to this story · Leave comment or trackback

Putting Skills to Good Work

I've just returned from a trip to Japan. While I was there, I had the opportunity to visit Kyoto Kakagu, a company that provide an excellent example of how to put skills to good use — and a nice contrast to the failure of Polaroid management's inability to save their company.

First, some background. If archaeologists dig up an old pot that's shattered, and they want to put it into a museum, someone has to reconstruct the pot and supply the missing pieces. When a museum wants to display model of an ancient building, someone has to create that model. And if a museum wants an exact replica of a valuable piece of sculpture, clay pot, fossil, meteorite, or document, someone needs to do that work. That was Kyoto Kakagu's business for many years. It's a fascinating spot to visit; I was quite jealous to see the workers handling clay pottery that was thousands of years old.

Here's the part that's relevant to this blog. A few years back, a customer approached Kyoto Kakagu with a request: could they please make a model of a human body for use in a medical school? Management had the foresight to realize that their skills in creating models for archeology could apply to medical models, and today Kyoto Kakagu supplies a wide variety of medical models. Small models of babies — ones with pulses and other medical responses — can train nurses how to correctly hold and care for infants. A human chest can be placed into a X-ray machine to emulate a patient with cancer in their lungs. Their latest full-size human model lets you listen to the heart, take blood pressure, check the eyes for proper pupil response, and record an EKG.

The lesson here is straightforward: your company's expertise may apply to other areas. In the book I discussed Polaroid, which despite a desperate need to find some way to keep your company relevant failed to find a modern use for their corporate skills as digital photography made the instant camera obsolete. Disaggregating your corporate skills from your current corporate business can open new horizons.

Topics:  · business · manufacturing · technology

Link to this story · Leave comment or trackback

Comments: 1, Trackbacks: 0

Cameras that Decide What Is Interesting

I ran across an interesting lesson for the world of business from the world of technology: In an attempt to solve a technology problem, a company called CSEM built a camera that decides what's interesting. I'm not quite certain, however, if it's really more the case that the engineers at CSEM learned a lesson from the world of business.

CSEM wants to incorporate cameras into cars. The goal is to provide safety features: Is the car veering off the road? Is another car too close for comfort? Traditional digital cameras simply take a picture and send all the data back to a computer, which then processes the picture to find other cars, the edges of the road, and similar information. Usually this requires reading data off the light-sensor chip inside the digital camera; and since there's a lot of information on that chip, reading the data takes a lot of time, and processing all that data takes even more time as well as an expensive computer.

But CSEM takes a different approach. First, they've built some clever intelligence into the light-sensor chip itself; secondly, the light-sensor chip can decide on its own what data are interesting. For example, if the car is attempting to decide if the car is too close to the edge of the road, the light-sensor determines which data are most relevant or "interesting" for the task and sends those data first; the computer receives and processes those data and can turn off the flow of less-important data. With this "interesting data first" approach less data are sent, which means that the whole process runs more quickly, and the pre-processing on the chip means less work for the regular computer — which can then be a less powerful, less-expensive computer.

Although businesses can certainly learn a lesson from this clever idea, I wonder if the engineers of this system were inspired by the business community. After all, the business community fueled the decades-long expansion in rapid package delivery: depending on customer needs, some goods must be delivered immediately while others can take longer. I can just imagine one engineer saying to another, "Maybe we can find a way to FedEx just the important data and send the rest later..."

Topics:  · technology

Link to this story · Leave comment or trackback

Amazon's Other Business

Amazon, the famous online bookstore, now offers a completely different set of services, ones that are geared not towards book buyers but towards Web developers. Amazon took its expertise in building large-scale reliable services for the World Wide Web and packaged it — amazingly affordably — for other people to use.

Let's say, for example, that I've developed a new online business that requires highly-accessible and reliable storage space. I can build a storage system myself, but that can be expensive and poses a huge barrier to entry. Amazon lowers that barrier — that's one of their stated design goals, to lower barriers to entry — by providing high-speed highly-reliable storage for rent. At a price of $0.10/gigabyte stored per month, and $0.20/gigabyte transferred per month, as an aspiring entrepreneur I can try something out to see if it's feasible without a huge up-front cost.

Or let's say I need a stack of 100 computers to test something, or perhaps I need this stack to run a web service. If I purchase these computers I must not only spend a few hundred dollars or more apiece, but I must also find an air-conditioned office with reliable Internet access, lots of electrical power, and a sercurity guard. Amazon will provide me with a single "virtual" computer, accessible online, for $0.10/hour. That stack of 100 computers would only cost $10/hour to run — testing a service or keeping it online becomes dirt cheap.

Amazon provides these services and more, including what they call "artificial artificial intelligence," a topic for a different day. Amazon has done an amazing job of disaggregating their business to resell some of their expertise, and even better they've done it in a way that has already jump-started dozens of new Web businesses.

Topics:  · business · design · software · technology

Link to this story · Leave comment or trackback

Expect Slow Sales of Next-Generation DVDs and Tough Year for Electronics Retailers

As the confusion continues in the long and pointless battle between Blu-Ray and HD-DVD (which I've written about before) over which standard will dominate next-generation DVD sales, the people who actually manufacture the players that go into consumers' living rooms will soon be able to easily support both standards.

As a result, smart purchasers — the ones who usually buy the latest and greatest and lead the pack — will likely sit on their hands and wait for the new players, which will not be available until 2007 at the earliest. Your local electronics retailer won't sell many of the available next-generation players, which is a bit tough on them. And since Microsoft Vista won't be ready until 2007, computer sales will also be slow until 2007, which will make things even more difficult for the retailers.

Both these incidents are sides of the same coin. Blu-ray and HD-DVD couldn't' find common ground, share authority, and settle on a single specification. The result is slow sales. Microsoft's Vista is tied so closely to the purchase of computers — Macintosh and Linux remain distant competitors in the consumer space — that any glitch at Microsoft severely affects sales. Lack of disaggregation has very real commercial consequences.

Topics:  · business · predictions · standards · technology

Link to this story · Leave comment or trackback

Comments: 2, Trackbacks: 0

Just Rent the Software, Dynamically

Back in my days at Dialogic, I suggested a novel approach to a hard problem. Let's say, for example, that you want to do speech recognition of someone who speaks Hungarian. Here in the US, there isn't much call for Hungarian-language speech recognition, and no one is going to install it and maintain it. So how can you do speech recognition in odd languages, or avoid the cost of installing any speech recognition, even English-language, in the first place?

I suggested that we disaggregate speech recognition; if you're in the speech industry, you may have heard me suggest it at a few talks I gave. We'd collect the voice here in the US and make arrangements with people overseas to provide recognition in their language; when we wanted a Hungarian-language recognition done, we'd connect to their overseas system and perform the recognition there. We'd get highly-specialized recognition that'd be simple to use — the usual benefits of disaggregation.

We never implemented this at Dialogic, but I've just found a company that implemented this "recognition rental" business model. They charge 750 € to purchase a recognizer, but if you like you can send data to them for recognition for 0.05 € for each successful recognition.

Topics:  · design · technology

Link to this story · Leave comment or trackback

Making Podcasts Useful

Podcasts provide information, but only if you're willing to listen to the podcast. Like many other people, I'm not willing to listen to a 15-minute podcast just to find out if it has information I need. I want the podcast to be searchable, just like a web page; if the phrase I'm looking for is in the podcast I will take the time to listen to the relevant section, and then perhaps the entire podcast — just like I do now with a web page.

And the search engine has to be very good. A web page takes only moments to skim after it loads into my browser, but audio is linear; I cannot listen to a few words here and there or skip to a summary paragraph at the end. Podcasts must overcome a huge burden if they are to become archival sources of web-based information.

Several companies offer search engines for podcasts using speech recognition technology. A new offering by Pluggd introduces a couple of new twists. First, Pluggd offers search suggestions based on your original search, and indicates where all these search terms appear in the file. These results provide context to the audio file and should help you determine whether the file you've found has the information you'll need.

Pluggd also says they'll disaggregate any user-created information from the audio file:

Users can click to listen to the file at that point, or select another option to tag, describe and share a particular section of the file. Castro says the company aims to set that data free, not keep it trapped in Pluggd.

This proposal takes Pluggd one step along the path to disaggregating information from audio, but as far as I can tell that's one step further than anyone else has taken.

Topics:  · technology · web

Link to this story · Leave comment or trackback

Is the Next Generation of DVD Players Doomed?

Let's say that every time you wanted to take a trip in your new Ford or Toyota automobile, you had to contact the manufacturer for permission: You would go to a web site and enter your starting point, destination, names of all passengers, and purpose for the trip. Certain trips would require a special, extra fee to the manufacturer if it decided that your trip wasn't on the list of "standard" uses for your car, or even if you gave too many rides to a neighbor who didn't have his own car. Oh, and good luck if the Internet was down, because you wouldn't be able to start your car — the manufacturers have to send an unlock code before each and every trip.

Not very attractive, right? In fact, if all new cars were made that way, you'd do everything you could to hold onto your present-day car and keep it running as long as possbile, and pray that the manufacturers regained some sanity before you were stuck with one of those new automobiles.

Sound ridiculous? Unfortunately, it describes exactly why I will hold onto my current DVD player as long as possible. Here's how the next generation of DVD players [Site requires free registration], HD-DVD and Blu-ray, will work:

If a consumer wants to make a copy of copyrighted content in his HD-DVD recorder, he first needs to hook his recorder to a network via Ethernet. Then, he goes to a Web site made available by a content owner, where he finds out how much it costs to make a copy in a certain resolution, such as high or standard definition. The consumer also has to inform the content owner of a destination medium (a portable media player, for example) where the copied content will eventually be played back.

I've written before about the incredible concepts behind Blu-ray; for example, the manufacturer can disable your player by remote control if they think you're using the player in ways they deem inappropriate. But I see the insanity isn't confined to Blu-ray, and that HD-DVD also contains these amazing "digital rights management" schemes.

The large "content providers" — the handful of music companies and movie studios that control the vast majority of entertainment in the US and worldwide — continue their campaign to redefine ownership. Digital rights management disaggregates ownership of "content" from the physical ownership of the data: I might own a DVD and the digital data burned onto it, but without a DVD player that has the correct cryptographic codes, I can't play the DVD and it's worthless. The large content providers are attempting to use that disaggregation to seize ownership and centralize it; they prefer that I purchase the DVD and then pay rent afterwards as well, a business model that brings them a steady revenue stream. Fortunately, current-generation DVD players won't go away soon, and the confusion in the marketplace between HD-DVD and Blu-ray will open up competition, and I can hope that competition will lead to more sane policies — ones that give me ownership of the content I purchase.

Topics:  · intellectual+property · standards · technology

Link to this story · Leave comment or trackback