- David Siegel
The World Wide Web has evolved from the original hypertext system envisioned by physicists to a planetwide medium that has already transformed most of our lives. What comes next should be truly stunning. Web 3.0 refers to a combination of advances that will change the Internet radically, making Web 2.0 look like a new paint job on an old car going down the Information highway.
In 2000 I gave a speech to business leaders telling them that as a futurist I could predict the future of eCommerce with 100% accuracy. The next phase of eCommerce, I said, was going to be called - commerce. There wasn't going to be anything special about purchasing goods and services online.
In this article, I'm not trying to predict the future of the Web. I'm helping create it. Web 3.0 starts now. The waiting is over. For a while, it's going to be called Web 3.0, to distinguish it from what we have today. After the transition phase, though, it's going to be called - everyday life.
[SIDEBAR: In 1996, Jon Bosak, Tim Bray, and others laid the foundation for Web 3.0 by creating XML, the eXtensible Markup Language, which began to add structure to the Web. Tim Berners Lee envisioned the Semantic Web as early as 1998. In my 1999 book, Futurize Your Enterprise, I described a set of commercial scenarios to help show how different this world could be. ]
I'll start by simply describing where we are today, but soon you'll see that the world of tomorrow is going to be very different.
The Web started as a way to share written electronic documents of all kinds. This blog is a Web 1.0 project. Amazon.com is a Web 1.0 project. Even Google (the search engine) is designed to figure out what you mean when you enter something like "Sofia Coppola" and show you an appropriate document. However, If a search for "Sofia Coppola" leads you to this page, you're not likely to be too interested in what you've found, because this page contains the words "Sofia Coppola" but is not about "Sofia Coppola". Web 1.0 generated a lot of words, pictures, sounds, documents, and MySpace pages (does Sofia Coppola have a MySpace page?). And all these pages are meant to be read (some of them enjoyed) by humans.
Web 1.0 is a great start, but its utility is starting to wane. We can continue to add more media (mobile, video, music, interaction), but it's getting harder to actually accomplish something. It's good for selling advertising. But it's far from turbo-charging your life or your business.
Do I even dare define Web 2.0, after the great Tim O'Reilly has told us in no uncertain terms?
Web 2.0 is a lot of things. It's Sarbanes-Oxley, which made it so only a few companies can go public, which led Google to go on a buying spree, which led VCs to invest in companies to sell to Google, creating a long slow boom in valuations, bringing in more investors, fueling more innovation. It's AJAX, technology that lets us run applications in our browsers rather than having to download a Java applet or a plug-in. Using AJAX, I typed this article right into my browser without opening Microsoft Word. Because of AJAX, I can play Sudoku on my cell phone without having to manage any downloads. It's a 10x reduction in cost of building and maintaining web sites and applications. It's cheaper bandwidth and servers, so you can now run enterprise-class servers out of your dorm room. It's mashups, which lets you scrape Craigslist and put all the puppies for sale with their Flickr photos on a Google Map. It's tagging, which brings us Del.icio.us and TagWorld, and video searching. It's Wikis, which let people contribute (and sometimes delete) content on a drive-by basis. It's collaboration, which lets people contribute to knowledge bases like Digg by acting in their own best interest. It's social networking, which has pretty much doubled the Web's traffic inside of 18 months and given us the power to flirt electronically. It's RSS and podcasting, which lets us subscribe to documents and have them delivered to us fresh off the hard disk.
Recently, IBM announced its "Marvel" system that can transcribe speech in a video to text you can read online. They call this exciting breakthrough (which several other startups have already developed) Web 3.0 - the Semantic Web. Is this Web 3.0? Is it semantic? No. Despite the hype, this is just transcription. It's still human-readable media in a more searchable format. Web 2.0 makes lots of things more searchable, but you're still searching for keywords and phrases and trying to infer context.
In short, Web 2.0 is really a faster/better/cheaper Web 1.0, with a few innovative concepts (and some rounded corners) helping glue things together. It's very cool. Almost everything Google does is Web 2.0. I'm even starting a few companies to take advantage of it (like Large Scale Dynamics and LTR.com), but Web 2.0 is still fundamentally Web 1.0 on steroids. It's faster and better, but it's not more meaningful.
Web 3.0 Overview
Whenever a new technology comes along, people first use it to recreate the old ways of doing things. The first photographs were still-lifes and portraits. The first films were of stage plays. The first typefaces looked like calligraphy. The first TV shows were based on radio shows. So it's no surprise that the Web recreated many familiar documents from the paper world: books, magazines, newspapers, TV shows, reports, radio shows, guides, movies, catalogs, contracts, and other human-readable documents. And we have indexes that help us find the words and phrases they contain.
The indexes try to be smart by implying meaning and context, but they have to guess what they think you are looking for. Ask Google to show you "a map of all the stores in San Francisco that carry maps of New York City." Google has a lot of the information, but Google has no way to make sense of what you're asking. Google doesn't look at the little contextual words "of" and "in", which in this case are very important. Although today's search engines get a lot of things right, they don't have the power to take us into the future.
The world of Web 3.0 is the world of semantics, where every piece of information has a standard meaning that can be understood not just by people but by programs designed to do mundane tasks for us. The standard example of Web 3.0 is scheduling. We all have calendars, but they exist just as text. Our calendars can't collaborate. When I try to arrange a dinner for 10 people, I have to manually coordinate the schedules of all the people and the available tables at the restaurants I'm willing to consider. That takes hours of going back and forth trying to maximize many variables in our heads and on paper. What I should be able to do is simply list my restaurants in order of preference, specify the preferred dates and times, ask all my friends to do the same, and let software agents work out all the scheduling details. A software agent is just a small program that carries out these kinds of tasks for us, saving the hassle and letting us make the decisions, rather than do the legwork. For that, all these systems need a common language that really doesn't exist. Yet.
Web 3.0 is much more than just semantics. Another example (of my own) is collaboration. Suppose I'm collaborating with a team of architects and designers on the design of a building. We decide to use X23 lights in all the corridors. The X23 lights (I'm just making them up) are energy efficient and light weight. By spacing them 8 feet apart, we would have to order 4080 lights. Today, we would look in the catalog or online, research the lights, then specify them by name and add their symbols into the drawings. Some one later would have to take our specification and turn it into an order - probably with mistakes after all the various changes in the plans. But if instead we could refer dynamically to their online specifications, then everything known about these lights would go into our design system, so the design system would instantly calculate how much weight they add to the entire building, how much light they would add to each floor, how much energy the building would use when they are on, create models of the lighting at different times of day, etc. And, if we changed to the X45 lights, which are heavier and more expensive but burn cooler and last longer, and space them 9 feet apart, we could see the impact on all these systems immediately. We could run comparisons, seeing how they light the hallways, how much the wiring has to change, how much the lights cost to maintain over time and where the savings are. Then, when we want to order, the system initiates and maintains the order automatically, so that each time we change the number of lights needed per floor, or the number of floors, the system automagically updates the order. As the building is built, the system changes from a design system to a construction system. Each light now has its own destination -- right on each light's box are the exact geospacial coordinates of where it's to be placed in the building. Eventually, the construction system turns into a maintenance system, and lights are re-ordered when they burn out (or even as they are predicted to burn-out, based on the actual data from other lights in the building and from lights in the same manufacturing run used by other buildings around the country. In this world, everything is not only connected, it speaks an interchangeable language of information, so each system knows what the other system knows.
That's Web 3.0.
When your doctor writes a prescription, why isn't the medicine simply waiting for you at your favorite drug store or shipped to your home? Why do you have to take a scribbled piece of paper to a drugstore and wait around while they fill it? Why does the pharmacist have to decipher the doctor's scribbles, hoping you don't have any drug interactions that kill 100,000 people in the US every year? Wouldn't it be better if we could catch problems the second the doctor tried to prescribe your pills?
The brakes on our car can pulse 30 times per second, but they can't start applying themselves early enough to prevent an accident. Our cars don't even know there are other cars on the road - that's so 20th Century! Our cameras are digital, but they don't know enough about the subject matter, conditions, or desired result to give us the photo we want. Our headphones are noise-canceling, but they can't help us hear our cell phone ringing. Laser scanners can add up our purchases at the check-out counter, but they can't count our calories or alert us when we buy products that have trace amounts of nuts someone could be allergic to. There's a lot of cool technology in our lives, but most of it is simply more, faster, and better at doing things the old way.
The best way to sum up Web 3.0 is to say that, in the future, everything will be smart. No longer will you need to go to several web sites to research flights and fares, then book your own ticket with your credit card information on a particular site. In Web 3.0, you'll just send your software agent on a mission to watch for and book a flight that meets your criteria at the best price (and at the last possible second, giving you more flexibility than you had before). You will be able to see the records for every flight you ever take on any airline, how many miles you have, where you went, what happened to your luggage, etc - all in one place.
We're going to need to get good at specifying the exact meanings of things, and we'll need to get good at expressing our research or purchase criteria more precisely, so various pieces of software can make sense of our requests. It's a lot easier than repeatedly trying to find the right phrase to type into Google!
Web 3.0 is to Web 1.0 as X is to Y. What are X and Y? That's a good riddle. It's challenging to try to come up with a good analogy, because almost nothing in our world is smart. Perhaps as we look closely at the three main areas of Web 3.0 we'll discover some useful analogies. The three areas are:
- Product and Service Descriptors
- The Online Desktop
Let's look at each area to see how the web will get a lot smarter.
The Semantic Web
I have a dream for the Web [in which computers] become capable of analyzing all the data on the Web – the content, links, and transactions between people and computers. A ‘Semantic Web’, which should make this possible, has yet to emerge, but when it does, the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines. The ‘intelligent agents’ people have touted for ages will finally materialize.
- Tim Berners Lee
He may not be Martin Luther King, but Tim Berners Lee does have vision, and the Semantic Web is a good one. If this smart web is going to become a reality in our lifetimes, there are two ways to build it: 1) use the human-readable text we see on the web today, and build super smart software agents that can magically make sense of everything in context, or 2) make all the information available in standard formats so that even relatively simple software agents can make sense of them. The first approach should take another 20-30 years of research into artificial intelligence. The second method should be able to scale up within about five years.
Academics and consortium members have been talking about the Semantic Web since before Tim Berners Lee wrote his May, 2001 Scientific American article describing several scenarios and the basic frameworks needed to bring them about.
Now, six years later, the Semantic Web is still a bunch of parts lying around in the shop, waiting for someone to put them together into something we can all use. We need a way to define an appointment, a price range, a contract, a ticket, an approval - in ways that software can understand and use on our behalf.
Here's my first analogy. The Semantic Web is to Web 1.0 as automatic scoring is to manual scoring in bowling. Remember the old days of bowling - Bowling 1.0? You got a big sheet of paper, you wrote down all the names, and then you'd make pencil marks on the paper and add up each person's score, according to the sometimes tricky bowling math. There was always someone like me in every group who volunteered to do the scoring, in much the same way we still write down golf scores today.
Then came computer technology. For a while, we had a computer screen where we could input each person's score and the computer would do the necessary math and display the result - Bowling 2.0. Today, the pinsetting machines tell the computer which pins remain after each frame, so the scoring software works automagically. Today, bowling alleys are smart. No longer do you have to worry about scoring - you just concentrate on your technique, and the system tells you the score, rather than vice-versa. Wow, that's pretty Web 3.0, don't you think?
It's a start. It would be even better if you could show up at the bowling alley, have the software recognize you, charge your account automatically for every frame, and then add your score for the day to your personal online data locker, where all your bowling scores live in one place. No matter where you bowl, every frame you throw will be recorded into your personal online bowling data locker, where you can chart your progress and use online tools to help manage your bowling life. You could even have a tournament against 10,000 other bowlers in hundreds of cities around the globe on the same day. That's Web 3.0. In the bowling realm, we're really not very far away.
Now think about golf. For this to happen in golf, we would need an enormous amount of technology that hasn't been invented yet. The smart golf club, smart golf ball, and smart golf course don't exist yet, much less talk to each other. Rather than having a scorecard that's a place to input your score, your smart golf scorecard of the future will tell you what your score is at all times and record your score in your personal online golf data locker, where you can chart your progress and use online tools to help manage your golf life. Today, in Golf 1.0, the best we can do is use a golf course's web site to sign up for a tee time.
Web 3.0 turns things around, from being a place where we use tools and applications (like reservations or ticketing) on different web sites to being a set of helpers that come to us, automatically assisting us in the context of what we are doing (everyday life). The Web has a long way to go until all the drudgery is done by software agents on our behalf. A lot of the pieces have been defined, but it's been slow in coming together. A search for "semantic web" shows many sites that haven't been updated in several years. Why? Why don't we have a Semantic Web browser? Why do we have PDF documents floating around, rather than semantic documents? Where is our utopian world of interconnected knowledge?
Let me see if I can answer that by way of example. I'm starting a new dating site, LTR.com. I think it's the best dating site ever built. To join my dating site, you need to fill out a profile that tells the system and other singles about you - your physical characteristics, your beliefs, your activities, your values, etc. All dating sites do this. You might ask - hey, isn't there a single form I can fill out and then send that description to ALL the dating sites?
Alas, no. It would be cool if we could have a basic identity and characteristics package that anyone could easily submit, and then each site with its own angle (one focuses on people who own dogs, another focuses on quizzes, another focuses on Long Term Relationships) could add its own special fields for you to fill out. But at least the basics would go in automatically, rather than having to re-invent the wheel every time. Why can't singles who have HIV see all the other HIV-positive singles in the entire world? It doesn't exist because there hasn't been enough demand to pull it through or enough push from the industry to agree on what it should be. And, in many cases, the businesses that get more customers on their proprietary platform make more money. This situation is typical of many industries.
Every time you go see a new doctor, you fill out three pages of medical history and emergency contacts, and they store it in some paper file somewhere, just like all the other doctors' offices do. How many of these do you have to fill out in one lifetime? Probably quite a few, because the medical informatics industry is very slow to adopt standards everyone can use.
Fortunately, World Vista, a new OpenSource standard is emerging for electronic patient records. It's based on Vista, developed by the VA, and now open to groups everywhere. It could form the DNA of a new world-wide system of interchangeable health information, and if everyone adopts the same standard, health care as we know it will change radically. By one estimate, 40% of all the medical tests performed in the US are done because previous test resultls are not readily available. According to Thomas Goetz, who writes for WIRED:
Using the VistA record system, the veterans department has managed to improve nearly every benchmark of quality in health care. In a decade, the department increased its pneumonia vaccination rate among at-risk patients to 94 percent from only 29 percent. That translates into 6,000 saved lives and $40 million saved each year from fewer pneumonia hospitalizations. On a host of other benchmarks — beta blocker use, cancer screening, cholesterol screening and so on — the department outperforms the nation’s best care.
Thanks to VistA, costs per patient at the Veterans Health Administration system are 32 percent lower, using inflation-adjusted dollars, than they were a decade ago. Over the same period, the medical consumer price index has increased 50 percent for the country as a whole.
There are a couple of other bright spots. As I noted in my 1999 book, colleges are starting to use the CommonApp. The Common App is a way for a prospective student to fill out one application and have it go to dozens of universities. It's been around for over ten years, yet only about 300 universities accept the common app. It's a start, but why isn't it simply accepted everywhere? The Common app just doesn't have enough demand from applicants or attention from school administrators to set a strong standard. It's the best we have, but it's far from a "smart" college application form.
Whether you're applying to a dating site, a new doctor, or a university, you need to fill out your name, address, birth date, contact info, and other essentials. And yet a common format for these small building blocks of information doesn't even exist. or, rather, they do exist, but no one is using them. The right way to build the Semantic Web is to use so-called microformats - small forms you can fill out once and then use as modules to create larger documents. And the Web we use in the future should be built of thousands of microformats that all plug-and-play with each other like lego building blocks, so you can construct most of a personals profile or college application or apply for a loan or rent a car simply by joining the lego blocks and reusing them as needed. Rather than reinventing the wheel every time you give out your information, all you need to do is keep track of the changes.
In Web 3.0, every single person has a personal profile online, just as most people have their resumes online. The search agents do the filtering and bring us results whenever they find anything. Dating sites turn into dating search agents. Employment sites turn into employment filtering tools.
I want to give my camera example, which comes out of my 1999 book, Futurize Your Enterprise. In this future world, most people's cameras will measure 0" x 0" x 0" and weigh zero pounds and zero ounces. Think you could carry one of those on your next trip? Today, in Camera 2.0, your camera has lots of memory and high resolution and you take photos and then upload them to your photo site, as I do with my art collection. In Camera 3.0 we realize that a) most people hate dealing with cameras, b) most people are terrible photographers, and c) most people just want good photos of themselves to magically appear on a web site (see "The Online Desktop," below) so they can look at them, share them with friends, and, even though I don't quite understand this behavior, print them onto paper from time to time. (Only a small percentage of people really enjoy screwing around with cameras and actually taking photos.)
How will this happen? Simple! Just add your Web 3.0 identity stack to something you're carrying on your person (a phone, purse, watch, or even a chip embedded in your skin), and turn on the switch that says "My ID number is xxxxx and I'm on vacation, I would like to be photographed." There are cameras everywhere. Most are now in the hands of people who enjoy taking photos. So, as you go do things, people take your photo and send them to your (anonymous) web site, where you can look them over and either accept them as gifts (if offered) or pay a few pennies (or whatever the photographer asks) for each photo you like. When you buy them, they go into your personal data locker (see below) or into your Flickr account, or wherever you like. Pretty much, all you need to do is get five people together in front of any monument, and ten opportunistic photographers will run up to take your photo. If five of you are in the photograph, you'll all come home to find the photos you were in waiting for you, along with links to the others' pages. In one day, you could get hundreds of photos taken of you by dozens of photographers you never even see. Fixed cameras sit at obvious points everywhere -- you take your own photo using your cell phone (or your headset) to send the camera a message saying you want it to take your photo and where to send the image.
If you don't want to be photographed, turn that switch off and no one will take any interest. Just turn it on when you want some one to photograph you. If you find this creepy, simply get a camera and bring it with you and don't forget to charge the batteries. Also, don't spend a week at Club Med, where they take your photo all day long and try to sell it to you that evening after dinner.
When people start to adopt these lego building blocks, the Semantic Web will take off like a rocket. You'll check into a hotel room in Dubai, and your bed will automatically be set at the firmness you like, the shower water will be at the temperature you like, your favorite TV show list will pop up on the TV screen, a list of invitations to events and restaurants in town you might be interested in will be shown in another window on the TV screen, and the alarm clock will automatically coordinate with your schedule, no matter how many times you make changes to it. You'll vote in elections from your cell phone. You'll get the care you need from any hospital in the world, because they will have immediate access to your vital health care records, even if you're unconscious. You'll be able to hand your big project off to someone else easily, so you can take on another project or go on vacation. Your phone will tell you when a call is important or "just checking in." You'll be able to make changes to the airplane you're designing, even as the first one is rolling down the production line. Buildings will reorder their own supplies. You'll never buy an album or even a single song again. You'll be able to switch all your accounts (cash, bonds, stocks) to another bank or brokerage using your browser in under a minute. You'll be able to see what your kids are up to at any moment, every mile you drive in your car, every photograph you take, and every agreement you sign. All via the Semantic Web.
So that's the second analogy. Web 3.0 is to Web 1.0 as Lego bricks are to sticks, glue, and cardboard. We just need people to start using those Lego pieces. And we need the security and validation structures to support those lego pieces, so we retain our privacy and can verify what others claim. There's a lot of work left to do, but various companies are working on pieces of the puzzle.
There's an interesting category of companies I call Web 2.5. These are companies that are trying to take an intermediate step, by categorizing the data that's either unstructured or loosely structured in a nonstandard way. eBay does this by asking people to categorize certain items themselves. They often don't do a good job, or the descriptions don't quite match up to the categories, but it's a start. Another company, called Silver Creek Systems, looks at all your data and tries to find common values it can create structure around. A company called Transparensee tries to make more sense of existing databases for singles and for housing. These aren't true lego blocks, because they don't allow much interchange with other systems, but they are an interesting temporary approach that shows the need for meaningful information.
This is the way things are heading. If we don't build interlocking Lego blocks, we'll have to get better and better at guessing. So we can either do it the easy way, or we can keep plowing money into artificial intelligence and sophisticated methods of guessing what we mean. With luck, the right standards will emerge, and Web 3.0 will wipe out the need for Web 2.5.
Universal Product and Service Descriptors
I've written about the product space extensively, both here and in my book. I want to paint a picture of how different the world will be when we have standardized structured metadata. What's metadata? It's descriptive information about a thing or a service. When you see a product description page at Dell.com describing a laptop, that's all metadata, including the photos, videos, and any sound files. A restaurant menu contains metadata for all the food they serve. In the last section, I mentioned your online storage locker for your information (You'll find the online data locker is a recurring theme in Web 3.0). In this section, I'll describe the various Descriptor Stacks and show how they will change our lives. Click this diagram to open it in a new window, then click on it again to enlarge it and keep it open as you continue reading ...
The first descriptor stack is our Identity Stack. It consists of the information found in all our government-issued identity documents, with varying degrees of protection and anonymity. There are online identities for individuals, couples, families, groups, corporations, and other organizations. It's beyond the scope of this article to describe how this stack works and how you keep personal information from falling into the wrong hands, but keep in mind that we do this with our bank accounts every day.
The Product Descriptor Stack is a new way of creating and using metadata. It consists of a base - the database of standardized product/service descriptions, a middle - the digital birth certificate and the personal data locker, and the top - the AJAX product/service presenter and the Semantic Browser. All these things combine to lay the foundation for Web 3.0 - ordinary life as it will be in the future.
At the bottom of the stack is a database of standardized product descriptions. In this database is the definitive description of any product ever made (are the Google people listening?). For example, there's a description of every make/model/year of every car ever made, starting with the Model A. The manufacturer's information is a small portion of the data available. Let's look at the 2012 Tesla 3, a car that happens to be near and dear to my heart. There is a huge amount of information on this car in the database: from the manufacturer, reviewers, magazines, insurance companies, enthusiasts, owners, etc. There are hundreds of photos, broken out by year, color, options, etc. Owners manuals are there, owners' tips, links to owners' groups, etc. Pretty much anything you want to know about this car is in this database, and it's easy to refer to this information from any application online. The key is that the data for every car is in a standard format, so you can compare all cars by any criteria you can think of: stated gas mileage, actual gas mileage that users get, repair records, weight, length, available colors, size of cup holders, crash-test ratings, BTUs of their heaters, etc. All products in any category can be compared head-to-head, apples-to-apples.
In another database, there's a standardized structured digital birth certificate for each car manufactured. The birth certificates have all the pertinent information for each particular car made, in addition to pointing to the generic description of the car (previous paragraph). In fact, each component of the car also has a digital birth certificate, so the car's digital birth certificate actually has a number of pointers to birth certificates at various suppliers. Does this remind you of the Lego building blocks we encountered earlier? The birth certificate never leaves the database, but information about who owns the car and what has happened to it gets added as the car goes from showroom to owner to next owner.
In the middle of the stack is my personal data locker, which houses my personal identity stack and an inventory of everything I own. It contains an ownership pointer that points to my car's digital birth certificate. Also associated is my personal owner's record, which includes every gallon of gas I've ever put into the car, GPS coordinates for every mile driven, all service actions, repairs, notices, condition of the car, and insurance records (remember, in Web 3.0, everything is smart). This is all under my control. If there's ever a recall on, say, a production run of 100 battery packs placed in that model of car and others like it, the company can follow the ownership pointers backwards to find the 100 cars that have those exact battery packs. It's important to distinguish between 1) the generic information for that make/model/year of car, which comes from the standard database; 2) the specific information about my vehicle - color, date made, options, photos, date/place sold - all part of the digital birth certificate, and 3) the owners' record - place bought, gas, mileage, condition, repair record, photos, etc. - all of which lives in my car's information folder in my personal data locker online. Still with me? Did you notice that all three places have photos? The standard description has photos of other cars of the same make/model/year, the birth certificate has photos of the car when it was made, and my personal data locker has photos of the car I've taken and various repair shops have taken over the years. Essentially, I have my own personal owner's manual in my data locker, and I can access it from my car or my phone any time I want.
At the top of the stack is the AJAX product presenter, which allows any web site to put up a window that lets other people view the car, according to the owner's permission. For example, I can list my car on eBay without moving any information. I simply refer to my car's folder in my personal data locker, and - automagically - people can see everything about my car that I want them to see. They can also see all they way down to the digital birth certificate and all the reviews and generic information, even down to the latest news on the company that makes the car. If I want to list the same car on Craigslist (yes, Craigslist still exists), I simply add the presenter to my listing on Craigslist and point it to my car as well. Note that I haven't copied any of this information, and if the car gets an oil change tomorrow, the people bidding on the car on eBay or looking at the car on Craigslist will see that reflected in the service record the day it happens.
(Note that when I say AJAX, I realize that in ten years it will be replaced by something else. I say AJAX now to prevent any confusion with a downloadable presenter. There will likely be several presenters using different technology. But that's fine, because each presenter uses the standardized metadata from lower down in the stack, so the actual technology for any particular layer can change easily.)
When I sell the car, all this information transfers to the new owner. The system tracks the full ownership chain and protects everyone's identity. I have more details on this, but I'm saving them for investors who want to help me make it a reality.
What if I've sold my car and want to buy a new mountain bike to replace it? Where would I look online for a mountain bike? Today, there are a hundred sites, from MTBR.com to Craigslist to eBay to Mountainbikeclassifieds.com to Pinkbike.com to Totalbike.com to BikeNashbar.com and many others. I have to find bikes that match my criteria by searching on keywords - good luck! Furthermore, my neighbor may have the perfect bike and want to sell it, but I have no idea. As I've noted in previous blog entries, searching by keyword is full of false positives (finding the wrong thing) and false negatives (what you're really looking for doesn't show up on search results).
To find a mountain bike, the Web 3.0 scenario is that I specify all the attributes I'm looking for - frame, transmission, wheels, suspension, price, etc. - using ranges or specific criteria, name any brand names, add my price and distance/shipping parameters, and then see all mountain bikes that fit my search criteria on a map, ranked by price or by condition or whatever criteria I want to see them ranked by. Can Google do that? Not last time I checked. If I don't see what I'm looking for, I want to just wait for my search bot to tell me a new bike has just come on the market that matches my criteria. Why should I have to go out searching all the time, re-running searches over and over looking for the right bike to show up in my neighborhood? Using Google to search for a bike, half the listings are from stores in England and prices are in pounds - what's up with that?
Why can't search be smart?
Search won't be smart until product descriptions are well structured and understood by all. The key to this new world of product descriptions, ownership, searching, and purchase is standardized structured metadata. It's so nonexistent that sites like Dell, c|net, eBay, and Amazon have had to invent their own taxonomies. Unfortunately, those taxonomies don't talk to each other, so general searches won't work because there are no standards. Have you ever tried to find a piece of clothing or a pair of shoes on eBay? Good luck! eBay actually has a size finder, and it helps, but there are still false positives and negatives, and there are no standards. You can't compare globes by diameter, cordless drills by voltage, stereo headphones with three drivers, or find all the American footballs (not soccer balls). Today, there are better online standards for avocado ripeness than for the most popular consumer goods categories.
Come to think of it, why are all the product descriptions locked up in web sites like Amazon.com, c|net, eBay, and millions of other databases? Why not just put the descriptions of everything you own and everything you want right on the Web and let the search bots find it? If everything were sitting out on the web, you'd be able to find everything, every single item that meets your search criteria. For example, if you're looking for a 17-85mm lens for a Canon camera for under $400, you could use a Web 3.0 search engine to see every single lens of this type, used or new, on a map of your city, and what prices people are willing to accept for them. If you found one nearby that someone wants $420 for, you might go for it, because you could go out and get it in a few minutes. Or you might go for the cheapest one on the map. If you had a lens like this, you could just leave it in your personal inventory locker, name your price, and wait for people to send you offers. This is called passive commerce, and it's a lot more convenient than checking a dozen sites every day or listing your item over and over, hoping the right buyer will just happen to see your auction or your listing that week. Rather than searching on dozens of sites, you could manage everything you have and everything you want right on one single site - your own private data locker. Why not let the information come to you?
In previous blogs, I've written that product descriptions are all written in different languages and standards - you can't make apples-to-apples comparisons from one web site to another if they all use nonstandard text to describe products. Recently, I've been searching for a pair of ice axes for ice climbing. It turns out that the descriptions, even within REI's web site, show the weight of various axes either in ounces, or in pounds and ounces, or in grams. You can't compare them apples-to-apples. As I've pointed out before, it's currently very difficult to learn the exact details you want on many products, much less make good comparisons.
Think of all the pages on the Internet that describe products. Whether it's an Amazon.com page, a page on Teslamotors.com, a page on Consumer Reports, Craigslist, or eBay, an Excel file, a Word file, or a product PDF, there are millions and millions of product descriptions online. Perhaps as much as 20% of the entire Web (without Myspace) is dedicated to product descriptors. How would you like to see product information? You may want different views depending on what you're looking for and what role you are in at the time. Someone doing research on plasma vs LCD displays is in a different mode from someone looking to buy a specific plasma display and bring it home and set it up that weekend. Comparing two houses or two diamonds side-by-side can't be done. Today, all these pages are made to imitate brochures or product data sheets or catalog pages or classified ad listings. In the future, they will be part of the semantic browsing experience, which is going to require a lot of infrastructure and a whole new way of looking at things.
The same is true of services. Have you tried finding services by keyword? It's pretty hit-or-miss. Suppose you have a new dating site, and you want to patent your amazing new matching algorithm. Try to find all the law firms that specialize in software patent law in and around San Francisco. Did you find them all on Google? I don't think you did. Today, the only way to find those law offices is to look for directories that have gathered the information. Directories are always out of date, even if they're online. And people compiling directories don't always look for the same things you look for. In the future, the law offices will have a chance to label their own sites with standard service descriptors, so their own web site tells the world what they do and the directories can syndicate that information, which will always be up to date. Today, most law offices have web sites today that list their practice areas, but this is just text. In the future, these descriptors will be standardized, so the search engines will be able to pick them up and categorize them appropriately. Then the search engine can compile lists of all the law firms who practice various types of software patenting, down to fairly minute detail (how many patents they apply for each year, average size of job, years doing patent work, number of attorneys, and so on). All service firms will take the small step of describing what they do using standard descriptors on their own web sites, and the directories and search engines will syndicate that information for all to use.
Describing goods and services requires a standardized ontology - a map of all items and attributes that everyone can agree on. An example of an ontology is the Dewey Decimal system in the library or the way products are arranged on the shelf at the grocery store. Every ontology has its drawbacks, but having a standard and extensible ontology for products and services will help form the base of Web 3.0. That's why the bottom of the stack has to be an open-standard ontology that everyone agrees to use.
When you try to categorize media, however, ontologies can get you into trouble quickly. Try organizing all your personal photos with a single ontology - it won't work. Try organizing your songs or your videos - the way you do it is going to be different from the way others do it. For that reason, tags are still the best way to add information that describes media. In Web 3.0, goods and services will use standardized structured descriptions with an ontology at the base, and media will use tags on top of standardized structured descriptions. For example, an album by the Scissor Sisters will contain factual information about the band, their producers, the recording, etc. But the genre and the categorization is best done by adding tags - words people associate with their type of music. Such tag-based ontologies are called folksonomies - people using tags build structures that can be useful, especially for describing media, even if incomplete or constantly changing. A folksonomy is probably the most useful way to index all the information found in blogs around the Web.
In Music 1.0 we have physical media, like disks and tapes. In Music 2.0, we have MP3 files flying around the web, people buying the rights to own a song, downloading it, then storing it to play it on their various devices. In Music 3.0, everything is streamed, and you just subscribe. You pay by the minute, and there's a scale of payment for various types of music, from free to expensive. You buy minutes in bulk and you simply listen to whatever you want as it streams into your earpiece or headphones or car stereo or home stereo. Same with all movies, although you'll be able to buy those per title as we do today. (Exception: new releases, which are still shown in theaters.) With media 3.0, you are always connected and always have access to the streaming media you want.
The presenter I've described goes beyond products and services. It shows schedules, tickets, weather data, information about your photos and other media, and many other things. The reason we have a presenter is that it can present information in human-readable form when you come to see it, but it can also send that same information to any program authorized to request it. The presenter could serve up the same information a hundred different ways, depending on the need and the context.
Is there a good analogy here to Web 1.0? I'm not sure there is. But let me give you a real-world example that gives us a ray of hope. There's one industry that's even ahead of the bowling industry, and that's publishing. Yes, good old-fashioned book publishers! The stodgy bespectacled people who sell paper artifacts have created a standard way to describe a book. It's called ONIX, and it's become an industry-wide standard. Today, publishers can maintain their catalogs using special software by GiantChair.com, and the software keeps all the ONIX descriptions for each book up to date on their web site. Then, rather than submitting that information t their distributor and having the distributor copy it and send it to retailers, the retailers can scrape their site using automatic agents, find every book description, and syndicate that information through their retail site. That's right - Amazon.com has software that takes updated book descriptions right from publishers' sites and promotes that information directly to the corresponding book page, so if the publisher makes a change to a book description, that change is reflected on the Amazon.com page for that book immediately. That's a good start!
ONIX provides a standardized generic description of a book title. In the future, you'll be able to put your entire library online via your personal data locker. You can describe each book, where you bought it, what condition it's in, and what price you want for it. If someone wants to buy (or borrow) one of your books and you're willing to sell it, you'll get an email offer and you can accept it. Rather than buying a book on Amazon.com, you might find that the book you're looking for is available for less right across the hall. ONIX is in place now, so to make this dream a reality, all we need is the online personal data locker and search bots that will do the drudge work for us.
Passive commerce is the future of commerce online. Whether you're General Motors or a kid with a comic book collection, you'll put your inventory online and let the search-bots connect you with buyers who have entered their search criteria. You won't go to a site to book a flight or a doctor's appointment. You won't go anywhere online, unless you're just browsing around for fun or trying to widen your search by getting new ideas. Instead, the Web will come to you.
The Online Desktop
The Web will come to you via the Online Desktop, which is your dashboard for interacting with the rest of the entire Web. Also called the WebTop, it's a web site where you keep all your information, dispatch software agents to do things for you, and see the results. As I've mentioned in previous blog entries, Google and Microsoft (and others) are going after the online desktop like they want to own it. What's the problem? They are both putting applications onto the online desktop that look and feel and work pretty much like our old desktop applications. This is to be expected, but it's far from Web 3.0. Done properly, Web 3.0 combines the Lego-like approach to information structures with the "always on" online desktop of the future that no one has really built yet. Let's break it down.
Single Source documents are key to Web 3.0. How many times have you received a document as an attachment? Once you receive it, it's dead. If you make changes and the author makes changes, it's difficult to reconcile the two files. If the author makes a small change, she sends it out again, and again, and again, so the attachments pile up and the emails and the versions get all mixed up. How many times has someone emailed you a photo with the file name of me.jpg? Once again, this is recreating the old world with new technology.
The Web 2.0 solution is Google Docs and Microsoft Live documents, which are traditional documents that live online. You don't need a word processor or a spreadsheet to open them, you just need a browser and the software comes to you. You don't send these documents to anyone. Instead, you send invitations to read or write these documents where they are. If someone has permission to write in a document you're editing, all the edits show up on both screens simultaneously. The document is always up to date. You can search the document's history for any changes you want to bring back, but there are no versions to maintain or get out of synch.
In Web 3.0, documents will have structure. Each kind of document -- a contract, friendly letter, business letter, medical chart, research report, product brochure, user's manual, menu, invitation - has its own inherent structure. And many of these documents lend themselves to an orderly building-block approach. It's easy to see the inherent structure in a menu: types of food, item names, descriptions, ingredients, prices, etc. Why use a word processor to build a menu, when you can preserve the structure and the meaning of all its contents using a specialized menu template and standardized menu descriptors? Now, each food item uses a standardized menu descriptor, and then I could send a query to a given restaurant asking to show which items are vegan, so I could order lunch delivered. Or, I could compare all the vegetarian Thalis from all the Indian restaurants in a given city, listed by price. I could see which restaurants in New York offer fresh live lobsters. I could find a place that serves vegetarian chili with cilantro. Or, rather, I could tell my software agent to do any of these things, and my agent would use my preferences to know what to do.
It's true that short stories and poems don't have much structure, and novels have even less. But many of today's documents are really just forms filled out. Even an invitation is really a structured document - a filled-out form. All information-based documents could benefit from the standardized structured approach, so that the same information I give to my doctor is used later to make my hospital bracelet, fill out my patient chart, fill prescriptions, schedule physical therapy, send me my x-rays, log the results of my blood tests, sign waivers, make claims against my insurance policy, make changes to my will, etc. The rule in Web 3.0 is: never duplicate information, always work from the master document.
In Futurize Your Enterprise I wrote an entire chapter on how Lawyers could use semantic building blocks to write contracts, so they could finally get paid for discovering facts and working with the law rather than tweaking words in poorly-written, haphazardly-structured documents. I'm not sure we'll see that kind of progress in my lifetime, but there are billions of dollars of QWERTY-style inefficiency in our economy waiting for Web 3.0 to streamline.
Web 3.0 documents will syndicate. Think of all the forms you fill out: magazine subscriptions, school documents, licenses, contacts, tickets, purchases, service contracts, donations to your favorite causes, checks, reservations, banking forms, etc. All these forms live in your online data locker. Now, if you change your address on the identity building block, every single document online would automagically update to show your new address - no extra effort needed! This is even better when you apply it to contacts - I can keep all my contacts online, and whenever any one of my contacts changes his/her address or phone number, that change goes to everyone's contact list automagically! You never have to send an email to everyone saying your email or phone number has changed! You can grab any cell phone and it shows your contacts in seconds, without copying anything. Make a notation to a number via your cell phone and you're actually changing the master online version. In fact, if you want to send me a package, you should simply address it to firstname.lastname@example.org and the Post Office's software will figure out (from my schedule) where I am at every hour of the day and intersect me with the package (first sending me a notice asking if I want to accept it or rerout it or send it back).
That's just contact information. Now think about catalogs. If you make a single change to a product description in your catalog, every catalog in the world will instantly show that change, because everything is linked, not copied. That's the power and leverage of Web 3.0.
The Online Data Locker is where you store all your information. You don't need to store anything on your computer. All your contacts live in one place, all your personal documents live in one place, even your drivers' license and passport will live here eventually. Think of ALL the documents you have in your paper files at home and on your computer - they will all live online some day. They'll be structured, they'll syndicate, they'll have protections, passwords, permissions, authors, viewers, and other properties. They will play with other documents. And you'll get at them from anywhere, any time.
Laptops go away. You won't have a laptop. Or, rather, any laptop is your laptop. You'll just grab any screen that's connected to the Internet, log in, and boom! - there's your desktop. You could use your earpiece for your phone to interact with your online desktop, listen to your messages, send messages, listen to music, etc. Your audio earpiece is fitted to your ear, so that's the only real piece of hardware you need. Your audio earpiece will get wireless music streamed in as you walk around town or drive down the street. Your audio earpiece talks to anything that can communicate with it. Keyboards, screens, and mice are everywhere and cost practically nothing. You don't need a hard disk anywhere. Nick Negroponte's hundred-dollar laptop is going to start looking awfully expensive in just a few short years!
Same with your phone. If you lose your phone, just grab mine and log in. In Web 3.0, phones don't have any brains or memory at all. They get everything from your online data locker and from the Net. They store all the music ever recorded, every movie ever made, and every photo ever taken, because they are always connected to the Web. By 2020, most phones will simply be a custom-fitted earpiece and nothing more.
The Semantic Browser helps you see all the semantic information in your online data in your particular context. If the AJAX presenter is pushing the information at you, then the semantic browser is the other half - a viewer you use to see the presenter in the way you want to see it. A semantic browser can make sense of all the metadata describing a product or service. For example, you can tell your semantic browser you prefer to see Metric measurements, you are looking for models rather than clothing, you're looking for specific phisical dimensions, you prefer to see reviews with higher rankings above company product information, you want to see time-sensitive information first, you prefer to buy things locally rather than pay for shipping, etc. You could be looking for stock photography of horses, not horses themselves; you could be looking for manufacturers, not retailers. Your semantic browser can tell all the presenters that you're in "research and compare" mode, not "compare and buy now" mode. A semantic browser would keep you from seeing all those ecommerce sites in the UK if you wanted to shop in the US (or vice-versa). You could see all the listings for a conference either by time, by category, by speaker - however works for you. There are a lot of interesting preferences to semantic browsing, and there will probably be different semantic browsers for different kinds of jobs in different sectors of the economy.
The data stays in one place and the applications roam the web. This is going to sound strange, but have you noticed that here in 2007 we are still moving data all over the place, from one application to another? I have Excel and you have Excel, so I send you a spreadsheet and you use your application to see the data I just sent. In Web 3.0, the data stays in one place, and the applications come to the data as needed. The document we're working on can require all kinds of extra capabilities as it evolves, and the software comes in from around the web to help us do what we're trying to do. There are no versions, no documents to leave behind on a particular machine. Doesn't this make more sense? It doesn't cost you anything to lose an application, but if you lose your data, you can be in big trouble! Why is all our data sitting on our personal computers, where it's vulnerable? Increasingly, more data is being stored online, but we still don't have the paradigm of applications moving to where the data is.
The software that helps an architect sketch a building is different from the software that creates the structural elements. Once the data is online, however, you can bring all different kinds of software in to test the structure, simulate seasons, design the window envelope, lay out the electrical system, design the water supply, heat and cool the building, maintain temperature, track maintenance orders, etc. All this software can now come to one place and work on the original document, rather than having to send updated copies all over and coordinate everything.
Let me now define three different types of tasklets: widgets, applets, and agents, all of which are small programs that do your bidding. (Not to be confused with bots and spiders, which search for information and compile indexes.)
A widget is a small piece of software that brings you live information that you can plug into any context you like. Most of us have weather widgets on our desktops already. Widgets will become much more useful in the context of a Web 3.0. document. You could use a widget to give you a list of all the countries in the world today, and if you use that widget, the list will automagically update itself whenever a new country is created. You could be writing a report on the financial markets and use a widget to show the performance of the S&P 500 in a box on the side of the page. The widget will always show the most current data.
An applet is a piece of software that comes to you and does your bidding on your data. An applet can come translate your document into Chinese. An applet can create a map of the trip you're on. An applet can help you manage your wine list. An applet can help lay out the electrical circuits on a blueprint for your house. An applet can do structural testing on the bicycle frame you're designing or look for viruses in your data or analyze CAT-scan data or help you plan a party. You could use an applet to record all your favorite channels and TV shows, and use that list to drive any device that can show a video signal. The applet can use an agent to be on the lookout for new material you might like.
An agent goes out and does things on your behalf. Your agent could look through many databases to find you a date for Friday night, a good restaurant in Budapest, a ticket to a sold-out concert, a video about ice climbing, a babysitter in Tokyo, the best croissants in Winnepeg, etc. Agents can interact and collaborate with each other to schedule meetings, agree on prices, negotiate terms, find a movie everyone hasn't already seen, etc.
This new world of tasklets will replace today's monolithic desktop applications. You could be working on a book to be printed, collaborating with 15 authors and 4 editors around the globe, working with their agents to schedule meetings and press checks, bring in an applet to manage the color calibration for the 3 different printing presses you're going to use around the world, use another applet to check the digital rights for all the photographs and dispatch agents to go get rights you don't own, use another applet to lay out the headlines for the chapter headings, and use a widget to include all the pertinent data that needs to be up-to-date before going to press. In this world, we no longer have monolithic applications, and we no longer have monolithic companies offering all the applets. Instead, we have a constantly improving ecosystem of tasklets competing for service niches that can be turned on or off easily.
You own your own data. In Web 3.0, electronic money doesn't live in a bank or a brokerage house. It lives in your online data locker just as securely but under your control. Want to invest in a mutual fund? Just bring the mutual fund to your money and give them access to it. Want to trade a stock? All the stock in your portfolio lives in digital stock certificates held by the exchange, or by a custodian. You simply own a pointer to each share of stock. (Moving pointers rather than digital stock certificates themselves is important, because then they all live in one place and all owners of one company's stock can be found immediately.) If you want to sell the stock, you can engage a broker to sell it for you and transfer cash to your account. You can let the brokers compete for each transaction. This is very different from today's world, where stocks and bonds are stuck inside a brokerage house and accounts are very difficult to transfer. Want to apply for a home loan? Your application is always filled out and updated, 24/7. Just allow the banks to look at it, tell them which property you want to buy, and wait for the offers to come in from all the banks that want your business.
The same is true for all your medical records. They live in your personal data locker under your control. Every x-ray, scan, test, visit, procedure, prescription, and appointment lives in your locker under your control. You decide who gets access to it. You can send a prescription to any pharmacy, and they instantly have access to all your relevant records. You can switch doctors or hospitals, or get a second opinion, without having to requisition copies of all your records.
The same is true for all your personal inventory. When you buy something, say a set of golf clubs, the ownership transfers to your personal locker the second you complete the electronic funds transfer to the retailer (or to another person). You can look at everything you own right in your personal locker from anywhere and get instant updates from manufacturers as necessary.
Your home has a smart owner's manual that lives in your personal data locker, and you can see at a glance how much energy you use, how much you spend on maintenance, the value of your home and everything in it, how much you could save by converting to different energy sources, how much energy you sell back to the grid, what needs replacing when, etc.
The Web is the Operating System of the Future. Come back in 20 years and it won't be Microsoft or Google. We won't have hard disks in our offices or living rooms. A set of open standards and full-tilt bandwidth makes this new world possible, and no one company will dominate our use of machines. Several companies will provide platforms for various kinds of documents in certain industries, but much of the really useful stuff will come from the edges. The future belongs to the small, the light, and the nimble. You'll forget that even your car is connected to the Web 24/7. Your data lives in one place, and the applications and applets come to you. You can fire them any time you like. We will look back on Microsoft Vista as the last of the dinosaurs before the impact of the Web 3.0 meteor that changed the climate and wiped them out.
The Semantic Web is smart, it's inside-out, it's made of interchangeable parts, and it updates instantly. Notice I have used the word "automagically" several times. This is funny. You'd have to be Rip Van Winkle to be amazed at today's bowling technology. We don't even think about the technology anymore - we just go bowling. At the moment, the term Web 3.0 helps describe how different things are going to be. But in 20 years, your kids won't think there's anything automagic about a grocery store cart that updates your total as you put items in or take them out and then charges your account on your way out the door. They'll just go shopping. Or golfing. Or do business without having to deal with the mechanics of moving the information from place to place. Your kids are going to call this newfangled Web "ordinary everyday life."
I want to be clear about something. To forge this world out of the keywords and documents sitting on the Web today will take a ton of artificial intelligence that we don't have and aren't likely to have soon. The horizon on "smart" AI systems seems to keep receeding as we race toward it. But the Semantic Web I'm describing isn't smart itself at all. It's simply well described. People take the care and attention to add a bit of information about everything they describe online using universal descriptors, and - voila! - the Web gets very smart very quickly. So we can either do it by hand, filling out online birth certificates and descriptors for everything and linking them together, or we can wait for the Web to evolve the smarts to figure things out for us. There is very little AI in the world I describe here, because it's a world in which everything is designed to fit together, and the context is specified by the user, not implied by the search engines. More accurate, less guessing - what more could you want in a World Wide Web?
Some day, companies will compete on the value they add to the economy, rather than trying to get a monopoly on the file formats or the distribution of information. Some day, the Internet playing field could be flat, rather than islands of information. This is the world we are tasked to create. The world everyone wants (except, perhaps, Steve Ballmer and Dick Cheney). If there's a Web 3.0 conference and all the key players get vaporized by a laser-guided bomb, it's not going to stop this future world from materializing. It's just a question of when. It's going to come from the margins into the center, so don't expect IBM's next press release to save the day.
There's a ton of work to do. We need more standards, more indistry buy-in, more public-accessible ontologies, and a new generation of tools. My goal is to raise a fund to start Web 3.0 companies here in New York City. If I had a partner in San Francisco, that would be even better. Any VCs want to join me in raising a fund together? I think the next Google-sized opportunity is here, somewhere on this page. I have an idea how to get it started (okay, I have a 28-page business plan), but I need help. I need the financing and partners to help make this world a reality. I'd also love to start having Web 3.0 conferences that can really address the marketing, business, and adoption issues. It's great to have a lot of the semantic layers in place, but getting consumers to adopt Web 3.0 means solving specific problems and addressing their needs. Commercializing Web 3.0 is where I want to contribute.
There are horizontal businesses everywhere here, from infrastructure to tools to services to applications to tasklets, soas, plug-ins, authentication, validation, identity, security, translation, and much more. I think they will be light and very competitive. There are vertical tools, systems, applications, and services to build in almost every sector. This is where the brands will be built. Done properly, Web 3.0 should rip the supply chain apart completely and reassemble it in a new and smarter way - completely integrated, and completely based on the syndication of standardized structured metadata. Done right, Web 3.0 will disintermediate industries that try to hang onto the means of distribution (classifieds, auctions, real-estate listings, stock trading, ticketing, etc) and create a new generation of value-add services and tools that will be about improving business rather than dominating the channel.
Hmm, come to think of it, I think I may have found the right analogy. Web 3.0 is to Web 1.0 as Web 1.0 was to the pre-internet world that only a few people can remember today.
I've been writing about this for a long time...
Futurize Your Enterprise, Wiley & Sons, 1999
Finally, I'm looking for a Web 3.0 intern to do some research for me and help maintain some lists of Web 3.0 developments. If you're interested, please send me your resume.
This article is brought to you by:
This is for keri dae sutterfield