Following in the footsteps of the great Ted Nelson, I have a hobby in which I sit in my living room writing white papers on the future. Back in 1991, I wrote a white paper on what I called The Communication Appliance, which I also called the computerless computer. This was before the Web. I wrote that there would be flat dumb terminals everywhere, allowing you to access all your stuff online. I said these terminals will be so cheap that they will be sent out as part of mass-advertising campaigns: it will be cheaper to produce the Spiegel and Sears catalogs by sending you a slate that plugs in and connects directly to the company's home page than to print and mail catalogs every quarter. The home page would be permanent, but you could still use the appliance to surf the rest of the sites. In other words, flat dumb terminals with keyboards, mice, voice input, printers, and other I/O devices will be everywhere. You would have 2-3 of these at home, one in your car, one in your office, and they would be ubiquitous so you could always just log in and see your desktop.
Welcome to 2006, where the race is on for the new Web 2.0 infrastructure, which will finally make the communication appliance possible. We're finally at the point where my vision might actually happen (although it's Amazon.com, not Sears, and I don’t even know if Spiegel is still in business). My goal here is to point the way to this future and perhaps put up a few signs for fellow travelers.
My basic assumptions are:
The Net is always on
Server space is cheap
Access is 24/7
Metcalfe’s law works only when people can interact
You’ll recall that Metcalfe’s law states that the utility of a network increases with the square of the number of nodes. The more people use online applications, the more the ecosphere broadens and people start building more and more useful applications on top of the base layer we are building now.
At this point, it seems like we have passed the event horizon of convergence. Things are starting to accelerate toward the center. Suddenly, everything is coverging toward some magic device that seems to be a combination of the best features of laptops, cell phones, computers, DVD players, MP3 players, video ipods, slates, watches, shoe-phones, and secret-decoder rings.
But as we will soon see, that vision is truly a look into a black hole.
Laptops Be Gone
Just as we managed to kill all the desktop machines and replace them with laptops, the next thing to do is kill all the laptops. I have a gorgeous sleek Toshiba laptop that is about 2cm thick, weighs less than a kilo, and packs easily into my book bag (note that books are still going strong). My laptop has a bright screen and excellent battery life. But what makes it mine is the pattern of bits on the hard disk. Since I’m storing more and more of my stuff online (all my photos are at www.dsiegelphoto.com), why not just put my desktop online and leave my laptop at home?
The laptop of tomorrow will
weigh nothing. It will be invisible.
Its dimensions will be 0x0x0.
Is that small enough for you?
You won’t need to bring it with you. It will already be there before you arrive. Wherever you go, there will be a dumb terminal that is so dumb it can’t even run Windows. This dumb terminal has no filing system, gets no viruses, and never needs to be backed up. It’s just a flat screen, some memory, a processor, and an assortment of peripherals. Your desktop is online and you access it from any connection, anywhere. Want to get some work done on an airplane flight, at a café, or in a hotel room? There will be a screen and a net connection waiting for you.
We won’t call them computers, we’ll call them connections. A connection has a screen, a keyboard, a mouse, probably voice input and sound output, and a 24/7 connection to the Net. It may have a printer or a scanner next to it, but most likely those devices will simply be sitting on a network, so you can use a net-based driver and designate your images to come from or go to any physical location you like.
In the future, you don’t go to the connection store and buy a connection, at least not the flat screen. It comes with your monthly bandwidth, the same as your digital router or your PVR that you get from the cable company. You can pay a few dollars more per month if you want more than one in your home.
The laptop of tomorrow will cost nothing.
Is that cheap enough for you?
Like the answering machine of the old days, your desktop PC will turn into a service. If you have a problem with your connection, your bandwidth provider will replace it. You may want to buy your own peripherals, but you can also go for the standard set of peripherals and just get a new mouse (or eye-tracker, or mind-tracker) sent to you whenever yours stops working.
In other words, as in so much of the rest of the economy, cost of goods is basically zero. The real value is in what you can do with it, and that’s a service you’ll pay for by the month.
Mobile Devices
You’ll need something on the go, so at the very least you can be rude to your fellow diners or detonate that bomb in your luggage. We all want to be connected, we hate to waste time standing in line, and we all want to be popular. So we’ll need something that’s a combination of our cell phone, email, video ipod, X-box, browser, and camera.
I think you’ll carry something that’s basically a communication device with an earpiece, a screen, and lens. Your earpiece(s) will be wireless. Your screen will be the smallest one you can carry that’s big enough to meet your needs, and eventually screens will unfurl or unfold.
What do people want to do on the go? They want to:
Talk with others in real-time (phone)
Listen to and view pre-recorded material (i-pod, porn, audible books)
Listen to and view live material (broadcasts, narrowcasts)
Send and receive messages (email, IM)
Query and get results from databases (i.e., directory assistance, traffic, weather, etc)
Play games (single player, multi-player)
Use specialized mobile applications (package delivery, sales, etc)
That’s the basic package. And, for most people, something similar to what we have today will work fine, but we’ll need broadband access to be everywhere, all the time.
Beyond that, you may want to capture the world around you and send it back to your server or share it with others. To capture images, your lens will either be really small and built in, or something bigger that you can use as a serious camera, or it will also include a microphone and serve as a video camera. It still uses the wireless device to send your images back to your server as you shoot, and the memory it needs will be on a chip, not on a hard disk.
I predict that laptops won’t get more mobile. I predict that phones will become much more mult-purpose, and that they won’t try to do what laptops can do today. They’ll simply help keep you connected when you are in between connections.
A World without Windows
The online desktop is already here. Several sites offer to store your files, from Yahoo (central server) to 247downloads.com (peer-to-peer), and you can get at them from anywhere you like. This is what virtual private networks are for.
Since your connection doesn’t have a hard disk, it has nothing to manage. It may have plenty of memory, so it can run several applications at once or deal with large data sets, but when you turn it off and boot it up it only knows one thing: the URL of the browser software it needs to load so it can start the browser. It then gives the browser the URL of its start page, and voila – you are looking at your desktop/start page. Your desktop keeps all your favorite bookmarks, documents, and applets ready to launch, so you can configure it however you like. It’s all customized and personalized to look the way you want, no matter how you access it.
The operating system really doesn’t exist, at least not on the connection. Instead, Firefox is the new desktop. The browser manages memory and communication back with the various servers, and the servers have their own operating systems.
The Online Office
It doesn’t take much imagination to build a word processor, a spreadsheet, or a database application online. These applications are a start, and there are plenty more coming. Google’s new database feature is a small step in the wrong direction. This is reminiscent of the first films being shot of stage plays – they simply imitated the old ways of doing things with new technology. But soon we’ll see that wiki applications – collaborative online applications – are beginning to flourish and make people much more productive.
Imagine a project-management application. Not only can the project manager modify the program, but all the participants in the project can contribute and change things constantly. Gone are the old days of single-user applications.
Many server-based applications, like Salesforce.com and Evite.com already exist. But the new protocols will allow you to keep everything on your virtual desktop and access everything with, I hope, some kind of similar interface. Ideally, they will all interoperate, so you can move information between programs on your desktop and between desktops with other people.
You may ask: “Okay, how will you do things like cut and paste?” Oy vey. Cut and paste is about as barbaric as e-mail. We can do better! I hope you won’t be able to cut and paste. I hope you’ll be able to move information much more effectively than cut and paste!
You may ask: "What about fonts?" Finally, fonts will be uniform resources, and you’ll pay a subscription to have the rights to use large families of fonts online. You’ll be able to get the rights to use an additional single font easily.
There’s nothing today’s operating system does that can’t be done better online. And we’ll be MUCH less susceptible to viruses. The sooner all this happens, the better. Here are some early examples of online applications ...
GoOffice -- A suite of office applications
Gliffy -- draw charts and diagrams
Writely -- a nice word processor
Zimbra -- calendar, collaboration, and messaging
Vertical Apps
Okay, you’ve tossed your laptop, gotten a computerless computer, and you’re connected. You can do all the word processing, spreadsheeting, and browsing you want. What about doing finite element analysis? Programming your automatic embroidery machine? Keeping track of your wine collection? Designing your next glider? Writing a screenplay? Configuring your insurance benefits?
All these vertical applications will work via the browser if they don’t already. The goal is for them to work together on your desktop, not in their own domains.
This is a strange concept, but let me propose it anyway: Today, the applications stay in one place and the data moves back and forth. In the networked future, the data stays in one place and the applications and viewers come and go. Is that crazy? It sure will lead to a lot less duplication of data. More and more, programs will become applets that work together with other programs so that we no longer need large applications that try to do everything.
Today, a real-estate management firm uses a custom application that has an entire back-office built in. But in the future, the back-office will be a horizontal application sitting on the Net, and the vertical (building-management) application will talk to it, whether you are using a small applet from your phone or sitting at your desk writing a proposal. Today, you must type your credit card information into every site you want to buy from. In the future, you’ll have a small bank application that manages your money and jumps in to help you purchase, categorizing what you buy and recording the transaction at the time you do it.
Metcalfe’s law continues. Web 2.0 will be much more productive, because some day, programs will talk to each other and we won’t have to re-invent the wheel every time we write an application. Vertical programs will be built on a strong foundation of other applications, so they’ll be able to focus on doing their special task even better.
Will you be able to make a drawing and do video editing online? Yes. You may have to pay for more bandwidth, but I’m convinced your online tools will be much more powerful than today’s desktop tools. Ten editors can cut a single film by leaving the film footage in one place and letting the editors work from anywhere - this is going to change how people work.
Here’s how I know: studies show that people tend to use about 5% of any application’s features most of the time, another 10% some of the time, and the other 85% either rarely or never. So why are we buying huge desktop programs – nay, huge suites of desktop programs – that have all the capability to do what everyone wants, when all we need is about 15% of what they can do? Have you ever tried to use the drawing tools in Microsoft Word? They’re a joke!! Why does Word have to try to solve that problem? Why not simply choose from thousands of little applets and build your own application, the way you want it, and add new tools as your needs change.
Have you noticed how many online applications have messaging, email, and storage of documents? Why do they all have to do those tasks separately? When all applications are on the Net, and when they can talk to each other, you’ll be able to add and subtract small bits of code that do exactly what you want them to do when you want them to do it.
Apple and Microsoft have been thinking about this for decades and doing the opposite. The new Microsoft Live project is no breakthrough, it's just another portal. It’s time to forget about the top-down approach to applications and just let a thousand flowers bloom online.
Social Networks
The big thing these days is social networks. There are thousands of them! So you have to get in the gate, then find your group, then link to people inside you have something in common with. And then you have to do the same thing in dozens of other social networks! I personally have profiles at Match, Meetic, Friendster, Tagworld, Partnerwinner, LinkedIn, and a dozen other sites.
Hello? Does anyone see something wrong here? Isn’t the web itself meant to be a social network? How about at least making these groups porous, so you can link to all your friends in each context just once? A good start is Tagworld.com, which gives access to tags that describe people within. Once you establish who all your friends or workmates are for a particular context (say a club or a team or a work group or a set of neighbors or a class), these new Web 2.0 applications should be able to share information among your group easily. But it won’t happen unless things change.
Why do all the social networks
not have applications?
Why do all the applications
not have social networks?
The answer is that both are too big. They don’t scale well, which is why they are so slow and bulky. This is what the World Wide Web itself is really for – rather than putting my profile up in 20 different places, I should simply put my profile on my own server and set up connections to other people around the Web. If Metcalfe’s Law is really going to start to work, we need to drop the big social networks and the big applications that try to do all things for all people.
A New Schema
There are plenty of blogs that talk about AJAX, XML, JAVA, and other technologies that are enabling Web 2.0. It’s a good start. With any luck, a set of interoperability standards will emerge to move us closer to the day when we can stop buying laptops. Here I want to talk about three big steps we can take toward interoperability: Standardized metadata, standardized user interfaces, and wikis.
As I mentioned in my last blog on structured metadata, a lot of people are talking about using XML to accomplish this and that. If we use XML, other web sites and applications will be able to know what our data means and will be able to make use of it. But if every system uses a different flavor of XML to describe the same thing, we won’t be able to interchange data without a set of dictionaries that translates from one set of XML tags to another.
I’m talking about metadata. It’s easy to describe a bottle of wine or a tractor or a pair of ice skates or a jet engine to your own system and perhaps those of a small number of vendors. But it’s another thing to create an industry-wide standard descriptor that everyone can use. Let’s look at two examples.
To describe a house, you need to specify the number of rooms, bathrooms, square feet, pool, car, garage, lighting, heating, etc. And you’d think there would be an industry-wide standard for doing that, starting with the schema behind www.realtor.com. But you would be wrong. Instead, we have a TOB situation (Tower of Babel), in which every site that lists homes for sale (eBay, Craigslist, owners.com, etc) has its own descriptor set. So even though you could map your data from one set to another with some degree of accuracy, in practice people don’t. In practice, these sites don’t communicate, and in practice there’s no real way to learn which homes are on the market in a given area without checking hundreds of web sites.
On the other hand, it’s easy to describe a book in a standard format. The book industry got together and created Onyx, an XML format for describing any book. This allows all the Web applications to pass book information back and forth easily. Publishers can use a system like www.giantchair.com to maintain all their book descriptions and then everyone who uses those descriptions will have accurate data to work with. All applications, from GooglePrint to Amazon.com to Alibris.com and others can work with the same data in the same format. Ideally, when the publisher makes a change to a descriptor field, that change immediately ripples throughout the online world, changing that item in every database. We’re not quite there yet, but starting with Onyx is a critical first step.
Unfortunately, most industries don’t have a standard descriptor for describing things, and the Tower of Babel just gets higher and higher. We need to have standard public descriptors for everything, from wine labels to engine parts to light bulbs to theater tickets to motorcycles to parachutes. When many things have standard descriptors, the new applications will start to show how seriously effective Metcalfe’s law can be.
The best effort we have in this area is called the Dublin Core
User interface standards need the same treatment. If we’re going to ask bits and pieces of code to operate in a standard way on standard data descriptors, then it would be great if the user interfaces were standard as well. A few people doing interesting work are:
Wikis rock. The more we can decentralize, the better. A wiki lets people collaborate by writing on the same page, rather than having to get someone else to publish for them. Wikis let me change the content of a web page simply by editing it. This is a great model for calendars, project-management systems, reminders, event-management, scientific collaboration, newsletters, etc. It blurs the line between newsletters and discussion groups, which can be a very effective way to collaborate.
As much as we love our webmasters, system administration, and tech support people, the world would be a better place if we didn’t need them in the first place. I claim that as soon as we have ubiquitous wireless broadband, we can start getting rid of all those people and get on to a more productive life. Wikis will be a big part of replacing the heavy-iron infrastructure with light, modular applications that make publishing as easy as creating. Eleven million bloggers (and rising) are the tip of the iceberg that will lead the way.
Not that all the technical people will go away. We’ll still need databases and database-driven systems to publish web sites. But much of what we call today’s internet and today’s intranet will, I hope, become much easier to use.
One of the keys to Web 2.0 is that building web sites and applications will cost a tiny fraction of what it did five years ago. Smaller applications and wikis are a big part of this. They eliminate all the experts in the middle and just let people communicate, have fun, create, and distribute their work. Here are some small wiki applications …
Kiko.com -- calendar
Jot.com -- a flexible tool for creating new applications
Basecamphq -- project management collaboration
Rememberthemilk.com -- to-do list and reminders
Bubbler.net -- Publish web sites wiki style
I’m not talking about all this because it gives my fingers something to do. We must have standard descriptors for things so we can build the wiki applets with common user-interfaces that talk with each other in Firefox so we can kill the laptops and we can have computerless computer connections everywhere and let the magic of Metcalfe’s law transform the Web into something thousands of times more useful than what we have today, which will keep us from straining our backs lugging all that generic hardware around and we can evolve as a species.
Note to Bill Gates:
thanks for the memories.
Please keep helping
kids in Africa.
The Next Macintosh should be no Macintosh at all. It should be Mac.com, where you can establish your space and put all your stuff. Get rid of the hardware, Steve, and let's build the online platform that brings in a new era of application development.
One of the great things about Nick Negroponte's $100 laptop is that it doesn't come with a hard disk. They wanted to include one, but it was too expensive. Thank god! Best thing that could have happened. Now the virtual desktop will have a chance to emerge and countries who buy the laptops will be forced to focus on providing 24/7 access to the Internet.
The list goes on. It's time to start the revolution.
Investment
We’re at the very beginning of Web 2.0, and I’m starting to become an optimist again. What I hope doesn’t happen is that the cool apps like Writely start showing up and, one by one, and get harvested and assimilated by big old heavy-iron companies as they ripen individually. The real benefit comes when they stay independent and all start to work together to enrich the ecosystem.
Investors like Steve Jurvetson and Tim Draper know that daring investments can pay off if you believe that you are making what people really want. I hope a few investors read this and it encourages them to bet on small things that can become ubiquitous. As investors, we shouldn’t be looking for the next big thing. We need to look for the next small thing: the best calendar, the best blogging tools, the best messaging system, the best micropayment system, the best applications that can work across the web, not just within a small community.
How can we help investors promote the growth of small useful tools? Let’s have some conferences centered around interchange and interoperability. Let’s help create the rich soil in which entrepreneurs can plant their seeds and let a thousand flowers bloom.
Summary
The Internet was meant to connect all the islands, but instead it created them. This is largely a result of using the new technology to simply recreate the old world online. Your desktop is an island. Your phone is an island. Your various accounts at eBay, Paypal, Google, Yahoo, Monster, Amazon, LinkedIn, Friendster, Shutterfly, Flickr, Winedemocracy, and all the rest are islands. Yes, they are connected but they aren’t interconnected. The data is still moving from place to place, rather than staying put and letting the applications come to it. And if we continue to follow the lead of Google, the islands will just continue to get bigger and bigger.
The playing field looks more like Swiss cheese than a football field. It’s no wonder people find the Internet frustrating! Search is at 5% of what it could do for us, applications are at about 5%, your computer’s desktop is a horseless-carriage concept from the last century, and your online experience does not become more productive with the square of the number of people using the Net.
We have yet to fulfill the promise of Metcalfe’s law. Instead, we have this barbaric thing called email, and 95% of our relevant information stagnates and is lost in the sea between the islands. Let’s start to learn to swim, rather than build more islands.
The one thing that still needs to fall into place is ubiquitous affordable mobile broadband access to the Net. When will we have this key piece of infrastructure? I think we’ll start to see it fairly soon, and it could roll out on a large scale within the next few years. Once that starts to happen, things will move quickly. With any luck, the last laptop will be sold in 2010. Let the revolution begin.
Here is a good list of Web2.0 software at the end of 2005.
