For the Week of June 30, 2003
The World Of Always On depends on one-chip applications connected by the Internet Protocol. But underneath every application you need an operating system, some standard way of making sense from the instructions in the application program.
Actually, you don't need an entire Operating System (OS). What you need is a kernel, a set of machine language instructions an OS is built on.
When I first started researching this book project, a few months ago, I assumed the answer was Tiny OS . Intel has based its own sensor networks on Tiny OS. It's tiny. And you can get GPS services through it , which is a key for applications that live in the real world. If your heart starts going kerblooey, a GPS-based system can tell the ambulance where to find you. The late Dave DeBusschere , Dick Schwarzlose , and thousands of others could have been saved had qualified help just reached them sooner. Faster notification of problems and immediate dispatch of trained professionals to the victim is the key to saving lives.
But Tiny OS is not the only possibility. If you used a kernel (the definition is like that found in nature, a germ, or the inner, soft part of a corn seed, where the genetic instructions are kept) you would have access to a big range of instructions, and the applications built on them. This would let you scale solutions, give you a ready user interface, and make everything compatible on a programming layer, not just a network layer.
Now here's a secret. Linus Torvalds didn't write the Linux operating system. What he wrote was a kernel, a Unix kernel, which was accepted for use in the General Public License (GNU) version of Unix. Torvalds wasn't as choosy about what might be done with his software as the GNU-folk, so he let his kernel go out under a variety of other licensing schemes as well. Linux is any version of Unix using Linus' kernel (which, by the way, has been revised and enhanced by a host of other people over the years).
Building chip-based applications on the Linux kernel is a great way to go. It's free, it's well understood, and it automatically interacts with the network applications of the Internet. Its history goes back nearly four decades, to the very first Unix operating system, built by Bell Labs.
IBM used Microsoft's MS-DOS for its first PC because Unix, at that time, was too bulky to be used with the slow chips then being made. MS-DOS wasn't a full-fledged operating system like Unix. It was just a "Disk Operating System," interfacing with the chip by putting complex instructions on magnetic storage.
As Moore's Law ground on this changed. Early versions of Windows were based on MS-DOS but over the last 10 years Microsoft has moved to a more Unix-like structure. (If you have a new PC you never see a black screen with text instructions on it as you boot-up, as I do on my old Windows 98 machine.) Today Windows is based, like Unix, on a kernel .
So developers of Always On applications have three choices for their single-chip applications - Tiny OS, the Linux kernel, and the Windows kernel. (If Apple had a Clue, they would have four.) The battle among these three is now taking place in what is called the Embedded Systems market.
If you thought the Copyright Wars were "Inside Baseball," with decisions made by a small, out-of-control elite, that's positively public compared to the Embedded Systems market. Embedded Systems lives in white papers , where engineers (with help from tech writers) try to educate other engineers about solutions to esoteric problems. Problems like, how the heck do I use this to do that?
The Embedded Systems market doesn't just represent a new frontier in the Microsoft-Linux competition. It also offers new opportunities for chip vendors to fight Intel, which now bestrides the world with solutions for either big OS. (Microsoft needs Intel now more than Intel needs Microsoft.)
In this Brave New World Microsoft has a big political problem. Many engineers hate Redmond. Microsoft also has a math problem. Anyone can read, understand, and write to the Linux kernel. Microsoft software is proprietary. It's like the old Steve Jackson game called Ogre (Steve was my first editor at the Rice Thresher in 1973-74) but with a difference. In this version the Ogre's opponent has an unlimited number of pieces to play. No matter how powerful the Ogre, it's still just Bill Gates. (Yeah - I know what you're thinking .)
So how can Microsoft escape the trap? Here are some random thoughts:
Here are a few things not to do:
- Take a page from Linus' book and license the Windows kernel under a variety of conditions. This does mean it won't remain a corporate secret, but (as we have observed) that is not entirely a bad thing.
- Create a VC fund that will back Always On applications. Don't put Microsoft's name on it, and make sure the resulting companies are independent.
- Create a separate company to handle licensing of the Windows kernel. Put its headquarters far from Redmond. Portland comes to mind, in part because Intel does so much of its development in Hillsboro, a Portland suburb.
- Don't try to create an "ingredient brand" like Intel Inside. That will only antagonize the market further.
- Work as closely as possible with Linux vendors to reduce compatibility problems, so products made with the Windows kernel interact naturally with those from Linux vendors.
- Create tools that link the Windows kernel to the Tiny OS.
In the end, the World Of Always On will have to take off on an Internet model. The transport protocol will be IP but more important, the chip-based products and software must be able to seamlessly interact with one another. By living with other software environments and nurturing its own developers (which it remains very good at) Microsoft can still take a hefty chunk of this market.
SSP (Shameless Self-Promotion)
Big news. I start work tomorrow as a business analyst with Progressive Strategies , a New York research firm that has the ear of the world's top technology companies. If we can get some of 'em to actually follow some Clues, it would be a very good thing indeed.
We have a winner. "The Blankenhorn Effect" has won the Computer/Internet category in the 2003 Independent Publisher (IPPY) awards . The book was also reviewed (favorably) by Ray Troxel of The Web Server Times last week.
"Dana, it is GOOD," raves Pete duPont, lawyer, futurist and once a candidate for President. "This is some really powerful 'stuff.' I think you've got a winner," says Drew Kaplan of DAK Catalog fame. One result is I have begun working on a follow-up book, describing the future direction of technology, to be called "The World Of Always On." Buy "The Blankenhorn Effect" at Amazon.Com , or at least say nice things. You can use the ASIN number, 1553953673, and recommend it to readers of other, similar books. You can also save on shipping when you buy the book at Amazon, over the cost of buying it elsewhere.
If you can convince some more published reviewers to read "The Blankenhorn Effect" and recommend it to their readers, please send me a name and address. In exchange, you'll get the PDF version of my second book, The Blankenhorn Effect: Boom, Bust & Beyond. This is a collection of columns from a-clue.com, organized chronologically and by subject, with additional commentary from yours truly.
You have my permission to forward this newsletter widely. And if you have trouble subscribing let me know . Remember: it's journalism that keeps the Clues coming...
Your list is your most important asset. But what happens when someone forgets who you are and you get on a "spam" blacklist? Your asset becomes worthless.
Need a-clue on how to avoid that? Get your list audited, and managed professionally, by the fine folks at Whitehat , part of the American Computer Group , a long-term leader in database services for direct marketers.
When your list is truly opt-in, not only do you become a white hat yourself, but your e-mails are read, even anticipated, by your audience. That means higher conversions and more money in your pocket.
If you're serious about Internet Commerce, you need Whitehat Interactive . Get it today.
Takes on the News
The biggest hurdle in front of an economic recovery isn't taxes or the deficit (those come later). It's not even tech policy (which still sucks).
The biggest problem is that crime pays. Catching Martha Stewart means nothing. She's like a street-level drug dealer. You might say the big players were Ken Lay of Enron, Bernie Ebbers of Worldcom, and Dennis Kozlowski of Tyco. But that's not all. It was the investment banks who created the crooked partnerships, who phonied up their advice to investors, who bankrolled the whole game. Relative to the crime they have already gotten off scot-free.
When this kind of financial abuse gets an all clear from the government, it's a short step to abusing the legal system, as SCO is doing. Or more precisely, Canopy Corp. , which has become the technology equivalent of the "professional victim," buying up assets solely for the lawsuits they might generate.
The SCO-IBM suit is already having a chilling effect on Linux development and implementation across corporate America. It is having no such effect in places like India, which sees Linux as a way toward gaining independence from the U.S., not just Microsoft.
This is part of the cost of the crime that such legal abuse represents. It goes to the heart of our industrial competitiveness. Foreign competitors aren't bound by the inefficiencies of our bureaucracy, nor of our courts. They will continue to bury us until we deal with that. Tort reform aimed at individual plaintiffs who have real injuries is what we are getting. Tort reform aimed at corporate abuse of the legal system is what we need.
Dan Gillmor had an epiphany recently after using a camera-equipped cell phone . This is changing the nature of photography, he wrote.
It's changing a lot more. If you're famous, walking through an airport, someone could be snapping your picture right now and sharing it. If you're not famous, and you're anywhere you shouldn't be, someone might just be snapping your picture and sharing it (perhaps with your loved ones).
The UK has been dealing with the impact of CCD cameras for some years now. They have solved a lot of crimes. They have also filled police files with more pictures than they can properly catalog. It's just one example of what happens when the police state decides it wants to know everything about everyone all the time - information overload.
But now this technology has become mobile and, with wireless networking, ubiquitous. What does this mean?
It means, as in many cases, that we need the protection of the law, and stronger protections for privacy. We can't afford to accept the excuse that these things "might" catch Osama Bin Laden. They won't. A warrant must be required before any picture is taken of anyone that is shared with police. We also need more protection against private police, who are frequently used by real police to do things they can't. (The Constitution doesn't really protect you against private action, only state action.)
If the rules for using this technology are determined in the present McCarthy-ite, religiously patriotic fever, we will be doing more than hurting innocent people - we'll be creating Luddites with a good cause. The continued growth of technology depends, absolutely, on the existence and enforcement of strong privacy protections, not only against Big Brother, but against Little Brother and Little Sister as well.
Gator Must Die
I first encountered Gator during the dot-boom. They were a nifty little company that made an electronic wallet. It filled out forms so you didn't have to. It wasn't Microsoft so people felt they could trust them.
During the bust, something changed. Gator first made some questionable deals, then some more questionable deals, and finally it fell completely down the slippery slope and became a bunch of hosers.
I'll give just one example. My son John is now 12, and he's a careful user. He has never been to the Gator site, and when offered something called "Precision Clock" he clicked "no." Although he has never given his name to any site that wasn't designed for children, he has lately been getting a lot of spam, even adult spam. He also has a firewall that warns him when any program on his PC tries to access the Internet, and just now he asked, "What is this Precision Clock thing?"
Sure enough, Zonealarm (his firewall) had spotted an attempt by this program to access the Internet and there below it was the author - Gator.
Dad had had enough. I went online and got him Spybot, a nifty little tool that tracks cookies and real nasties, then gets rid of them. A scan revealed 255 suspicious entries, 173 of which Spybot felt were serious. Gator (and programs associated with it) represented nearly one-third of that total.
Out they went. They should be gone from your computer, too. If Gator asks to advertise on your site, just say no. Anyone who does take their money, I pity them. You piss off a dad, and you've pissed away a lot, and Gator, this dad is plenty pissed.
Clued-in is the decision by the National Association of Security Dealers (NASD) requiring members to archive Instant Messages for three years . Remember that this only applies to brokers, no one else. Anything that forces those people to be honest brokers is a very good thing. (Yes, dishonest brokers can hose people through private accounts, but if just one victim of such an IM saves the message and delivers it in a complaint, the broker's career is over.)
Clueless is the U.S. decision to start censoring the Iraqi press . Transforming liberation into occupation turns World War II into Vietnam faster than you can say "six Shiites saw Saddam sink slowly in seal shit."
A-Clue.Com is a free email publication, registered with the U.S. Copyright
Office as number TXu 888-819. We're on the Web at http://www.a-clue.com.