Like the Volkswagen Beetle and modern freeway systems, the Telex messaging network comes out of the early period of Germany’s Third Reich. Telex starts as a way to distribute military messages, but soon becomes a world-wide network of both official and commercial text messaging that will persist in some countries into the 2000s. Telex uses teleprinters, which date back to the 1910s for use in telegraphy. But instead of using pricey dedicated telegraph lines, the telex system connects those teleprinters to each other over voice telephone lines, routed by modified telephone switches. Wireless versions of Telex soon connect remote regions of the developing world.

Belgian Paul Otlet has a modest goal: collect, organize, and share all the world’s knowledge. Otlet had co-created a massive “search engine” starting in the early 1900s. His Mundaneum now combines enhanced card catalogs with sixteen million entries, photos, documents, microfilm, and more. He is working on integrating telegraphy and multiple media, from sound recordings to television. In the 1930s British writer H.G. Wells and American scientist Vannevar Bush are advancing similar goals—Wells with his “World Brain” writings and Bush with the Memex, a sort of microfilm-based Web browser. These approaches to organizing information differ. But all share key features of today’s Web, including automated cross-references – which we call hyperlinks.

With side-by-side screens, the imaginary Memex desk is meant to let a user compare and create links between microfilm documents, somewhat like today’s clickable Web links and bookmarks. The idea is that people will continually build on each other's associative trails through the world's knowledge, helping tackle the growing problem of information overload. The Memex is the brainchild of top U.S. scientist Vannevar Bush, an analog computing pioneer who had helped oversee development of the atomic bomb. The basic mechanism he suggests is a microfilm automatic selector similar to those built by optics pioneer Emmanuel Goldberg in the early 1930s. Bush publicizes the Memex concept in 1945 articles in The Atlantic Monthly and Life.

Computers “talk” over ordinary voice phone lines through modems. Developed in 1949 for transmitting radar signals by Jack Harrington’s group at the Air Force Cambridge Research Center (AFCRC) near Boston, the modem modulates digital data into sounds, and demodulates received sounds into digital data. (MODulation + DEModulation = MODEM). Modems will be adapted to computers in 1953 for the upcoming SAGE system, and commercialized by Bell Telephone in 1958. By letting computers use normal voice telephone lines, they offer greater coverage and lower costs than dedicated telegraph or leased data lines.

Phone companies develop digital transmission for internal uses – specifically to put more calls on each of the main lines connecting their own switching centers. By 1958, this produces the T1 standard still used in North America. By the 1980s, phone companies will be leasing digital lines to commercial customers.

Designed to detect Russian nuclear bombers, the IBM-built SAGE (Semi-Automatic Ground Environment) pioneers many technologies including a special-purpose form of networking. There are 23 computer centers across North America, communicating with radar stations, counter-attack aircraft, and each other — all in real-time, as potentially threatening events are happening.

Besides networking SAGE also helps pioneer interactive computing and multi-user systems. Hundreds of people use the system simultaneously, interacting through groundbreaking graphical consoles. Each console has its own large screen, pointing device (a light gun), a telephone, and an ashtray. Ever on the alert for a Soviet attack, SAGE operators would describe the experience as endless hours of boredom…broken by seconds of sheer terror. By the early 1960s, intercontinental ballistic missiles (ICBMs) will make the bomber threat – and SAGE itself – partly obsolete. But the system will operate until the 1980s.

In the 1950s several visionaries including Ted Nelson and Douglas Engelbart independently suggest computerizing the concept of cross-references, creating the clickable link we use on the Web. Nelson calls it a “hyperlink,” and the computerized text “hypertext.” Along with graphics pioneer Andries van Dam they develop many core computing functions such as word processing, online collaboration, and hypertext links. J.C.R. Licklider’s 1960 book Libraries of the Future outlines a related vision, which adds mild artificial intelligence to the mix.

By the early 1960s many people can share a single computer, using terminals (often repurposed teleprinters) to log in over phone lines. These timesharing computers are like central hubs with spokes radiating to individual users. Although the computers generally can't connect to each other, these are the first common multi-user systems, with dozens of people online at the same time. As a result, timesharing pioneers many features of later networks, from file sharing to e-mail and chat. Typical 1960s users are a mix of business people, bank employees, students and researchers, and military personnel.

Some of the first general time-sharing systems are CTSS at MIT, and PLATO II at the University of Illinois.

Online transaction processing makes its debut in IBM´s SABRE reservation system, set up for American Airlines. Using telephone lines, SABRE links 2,000 terminals in 65 cities to a pair of IBM 7090 computers, delivering data on any flight in less than three seconds. It remains the basis of modern travel sites including Travelocity. SABRE has been adapted from IBM's earlier work on SAGE.

Used by Texas oilmen, the Carterfone acoustically connects mobile radios to the telephone network. Telephone companies sue in 1966. The FCC supports Carter, freeing U.S. telephone lines for many uses—including later answering machines, faxes and modems. Users in some countries will wait until the 1990s for similar freedoms. Modems create a kind of de facto net neutrality; telephone companies have no control over what is sent over their lines with a modem.

Since telegraph days, people have been refining techniques for squeezing more connections onto a single wire. Early multiplexers for computers let up to 15 terminals share the same line, by assigning each of them a particular frequency (Frequency Division Multiplexing). In 1968, a new generation of time-division multiplexers gives each user a small slice of time in turn, radically expanding the number of computer terminals that can share the same line – from 15 to 45. This dramatically lowers costs for businesses, and the multiplexer market becomes highly competitive. Advances in acoustic couplers – the devices that let you connect your terminal to a regular telephone handset – also help more and more people go online.

Douglas Engelbart and his team at SRI, with funding from ARPA, unveil their experimental ‘oNline System’ at a computing conference in San Francisco in what will become known as the ‘mother of all demos.’

For ninety minutes, the stunned audience of around a thousand witnesses collaborative editing, videoconferencing, word processing, and a strange pointing device jokingly referred to as a "mouse." Every paragraph or other element in the NLS system can be linked to other elements using hypertext links.

Across the country, visionary Ted Nelson is working with Andy van Dam to create another early hypertext system, HES (Hypertext Editing System).

Switched on in late October 1969, the ARPAnet is the first large-scale, general-purpose computer network to connect different kinds of computers together. But others come online within weeks or months. 1969-70 marks the start of Britain’s NPL network, the wireless and more specialized ALOHANET in Hawaii (also ARPA funded), and the HLN (High Level Network) for the SITA consortium of commercial airlines. Work begins on France’s CYCLADES network not long after. The ARPAnet’s massive funding will help it pull ahead of rivals.

The ERMA system had revolutionized behind-the-scenes check processing in the 1950s, spawning the funny letters still at the bottom of checks today. During the 1960s researchers in various countries have been working on bringing automation – and online transactions – to customers in the form of an Automated Teller Machine (ATM). Barclay’s Bank in the UK has likely been the first to put one in operation, in 1967. By decade’s end many systems are up or being planned in Europe and North America. The paper used by some of the first ATMs is slightly radioactive, to be machine readable. The 1970s will also see rapid growth in behind-the-scenes financial transaction networks, like SWIFT for wire transfers.

By1971 Sam Fedida at the British Post Office and teams at the BBC and the IBA (Independent Broadcasting Association) have started developing Web-like information systems that use televisions for display. The two latter systems, based on work by Philips, broadcast data on an unused portion of the TV signal. They evolve into the Teletext information services found on most European TVs into the 2000s. Sam Fedida's videotex standard at the Post Office (which also runs the telephone system) uses phone lines, and has high ambitions for broad-reaching uses like today’s Web. It becomes the foundation of England’s Prestel and, later, France’s wildly successful Minitel.

In the early 1970s email makes the jump from timesharing systems – each with perhaps a couple of hundred users – to the newly burgeoning computer networks. Suddenly, messages are free to travel anywhere the network goes, and email explodes. Ray Tomlinson of Bolt, Beranek and Newman chooses the now-iconic “@” sign for his networked email protocol on the ARPAnet and by 1973, well over 50% of traffic on that research-oriented network is email. Nearly all other networks add email features. On the PLATO educational system, the email features of PLATO Notes are a runaway success both for person-to-person mails and as the basis for discussion boards.

In 1973, ARPA funds the outfitting of a packet radio research van at SRI to develop standards for a Packet Radio Network (PRNET). As the unmarked van drives through the San Francisco Bay Area, stuffed full of hackers and sometimes uniformed generals, it is pioneering wireless, packet-switched digital networks, including the kind your mobile phone uses today. A related set of experiments test out Voice Over IP (like the later Skype). The van will also play a huge role in 1977 as a major birthplace of the Internet.

Early networks successfully connected computers. But different kinds of networks couldn’t link to each other, limiting the size of online communities. So, the next challenge has been creating “networks of networks,” a process called internetworking or internetting.

France’s CYCLADES and Britain’s NPL network are experimenting with internetworking by 1973 with the European Informatics Network (EIN). Xerox PARC begins linking Ethernets with other networks using its PUP (PARC Universal Packet) protocol the same year. Both these efforts will influence the development of ARPA’s TCP/IP internetworking protocol, first sketched out in 1973 by Vint Cerf and Bob Kahn. ARPA has a practical need to link its original ARPAnet with its newer networks like the Packet Radio Network (PRNET) and Satellite Network (SATNET). In 1977 Cerf and Kahn will link the three networks and prove the efficacy of their TCP/IP protocol in a dramatic round-the-world transmission from a moving vehicle, the SRI Packet Radio Research van.

Computers have to communicate down the hall, as well as globally. Local area networks (LANs) evolved from the early links to peripheral devices such as terminals and printers. 1973 marks the birth of the standard that will eventually prevail: Ethernet. Created as part of Xerox PARC’s sweeping vision of an “office of the future” comprised of connected PCs, Ethernet adapts techniques from the wireless ALOHAnet to treat cables as a passive medium, like the air (“ether”) between radio stations. But it will have stiff competition from various local network standards including IBM’s formidable Token Ring and Datapoint’s ARCNET.

IBM has been building hierarchical, special-purpose networks since the SAGE system in the late 1950s and SABRE not long after. In 1974 it announces Systems Network Architecture (SNA), a set of protocols designed for less centralized networks. SNA will evolve into an internet-like network of networks, albeit one reserved for those that were SNA compliant. DEC and Xerox will also begin commercializing their own proprietary networks, DECNET and XNS. At it’s peak around 1990, IBM’s SNA will quietly carry most of the world's networking traffic.

The mid-1970s brings a number of commercial networks for corporate customers and professionals to choose from. Most are packet-switched, like Telenet, Tymnet, and other major players. The ill-fated Datran and a couple of others are circuit switched like a telephone system. Six year old CompuServe is also becoming a major supplier of corporate network services; it’s consumer-oriented CompuServe Information Service will follow at the end of the 1970s.

Anyone can walk up and use this terminal, connected to a timeshared mainframe computer, for posting messages and announcements. That's a radical idea when computers are mostly inaccessible to ordinary people, and seen by the counterculture as tools of government and corporate power.

Queen Elizabeth II of the United Kingdom sends out an e-mail on March 26 from the Royal Signals and Radar Establishment (RSRE) in Malvern as a part of a demonstration of networking technology.

Until the late 1970s the momentum in computing has been all about togetherness – users first sharing computers, then linking over networks and soon networks of networks. But the rise of the personal computer from the mid 1970s makes something once unthinkable an everyday reality: a standalone computer for just one person. While the new machines can be connected to networks and to each other, a lot of users both at home and work don't bother. They run their own programs off of floppy disks. The “personal computer revolution” begins to push back against the centralized control of network system administrators, a trend that won't fully reverse until the 2000s and the emergence of “the cloud.”

The first Multi-User Domain (or Dungeon), MUD1, goes on-line. Richard Bartle and Roy Trubshaw, two students at the University of Essex, write a program that allows many people to play against each other on-line. MUDs become popular with college students as a means of adventure gaming and for socializing. By 1984, there will be more than 100 active MUDs and variants around the world.

Personal computers have started to slowly take off in North America by the end of the 1970s, a decade earlier than most other parts of the world. Connecting them to remote servers can be a nightmare of endless settings and false starts, accompanied by the squawks and squeals of an expensive, finicky modem. Most computer owners don't bother, but by 1979 a subset of brave or stubborn ones are subscribing to early online services like MicroNet (later CompuServe Information Service) and The Source, or connecting to Bulletin Board Services (BBSs) hosted on somebody else’s minicomputer or PC. By 1990 more than two million North Americans will be online for discussion groups, shopping, news, chat, e-mail, and more; the early online services have been joined by AOL, Prodigy, and others. This dial-up world pioneers much of what we do on the Web, though in a more communal setting.

From the late 1970s on academics and geeks continue expanding “techie” online communities like Usenet (a message board conceived by Duke University students) and BITNET (a network for file and email exchange). One of the most durable online communities, Usenet provides topic-oriented “newsgroups” for collaborative discussion, and its community and ethos will shape the early Web.

As networked computers arrive in offices through the 1970s and 1980s, professional information systems continue to blossom. LEXIS (which has roots in the computer utilities of the 1960s) provides access to legal cases. NEXIS adds a massive searchable database of news articles. There are a number of industrial purchasing systems based on “Electronic Data Interchange” standards for computerized transactions. DIALOG provides pricey information for businesses, and dozens of more specialized services address particular niches.

John Shoch and Jon Hupp at the Xerox Palo Alto Research Center create the computer "worm," a short program that searches a network for idle processors. Initially designed to provide more efficient use of computers and for testing, the worm has the unintended effect of invading networked computers, creating a security threat. Shoch took the term "worm" from the 1976 book The Shockwave Rider, by John Brunner, in which an omnipotent "tapeworm" program runs loose through a network of computers.

In 1980 Tim Berners-Lee at the CERN physics laboratory creates Enquire, a networked hypertext system used for project management but with far greater ambitions. It seeks to categorize hyperlinks in a way that can be read by computers as well as people. He later claims he hadn't been aware of earlier hypertext work at the time, so it may be an independent reinvention. He names the program after a Victorian advice book and encyclopedia he had loved as a child, *Enquire Within (about Everything)." Berners-Lee will go on to invent the World Wide Web, partly based on Enquire.

Free! That’s often an effective way to attract customers. In 1981, France Telecom offers free Minitel terminals to every phone subscriber, launching the first mass “Web.” Minitel will have tens of millions of users by 1990 and online services such as newspapers, train schedules, tax filing, and erotic classified ads as well as email and chat. The ‘80s Minitel boom heavily foreshadows the dot-com boom. But the business model is different. Customers pay by the minute for access to Minitel services (sites), charged on their phone bills; France Telecom keeps about a third and passes on the rest to the service provider. As in the later Web, Minitel service providers run their own servers. But they also pay France Telecom a fee to connect to its network. Despite major efforts in the US, Canada, and Europe, similar videotex systems will fizzle outside France.

Protocols like Ethernet or Token Ring have established low-level links between computers and peripherals in the office. But that's only part of the solution – workers still need to do higher-level tasks such as sending e-mail, exchanging files, and sharing printers.

This need yields a hodge-podge of third party “network operating systems,” including Novell Netware, and built-in solutions like Apple’s AppleTalk. In the 1990s, Internet protocols will replace them all.

Bolt Beranek and Newman, which had built the original IMP and designed important parts of the ARPAnet, had also been a key participant in ARPA’s 1977 internetworking experiments. They produce early switches like the C/30 Communications Processors, but nimbler rivals like Cisco will soon overtake them.

For music, the CD (Compact Disk) had come out in 1981. For data, the big date is 1984 when Phillips introduces the CD-ROM (the last part stands for Read Only Memory). In the late ‘80s and early ‘90s, affordable computer CD-ROM drives will enable multimedia encyclopedias, games, novels and reference information on CD, often with hypertext links for navigation. This content will turn out to be basically ready-made for the Web, and many publishers will simply port it over directly. For content that uses a lot of video or images, CD-ROMs will remain an important adjunct to Web browsing throughout the early Web era.

Many agreed on the goal: to develop a global network of networks, or an “internet” in the parlance of the time. They don't agree on how. By the early 1980s, several different national and corporate protocols are competing with each other. OSI (Open Systems Interconnect) is the first with international backing, and support from the International Standards Organization as an official standard. Begun in 1977 by a member of a team that pioneered internetworking on the French CYCLADES network, OSI is officially published in 1984.

Digital Equipment’s DECNET is a strong competitor. IBM’s System Network Architecture (SNA) dominates the world of corporate computing, and will carry the majority of the world’s networking traffic up through the late 1980s. The dark horse contender is ARPA’s Internet Protocols (TCP/IP), defined only by a self-governing community of those who have access to this closed network, mostly U.S. military and computer science researchers.

Stewart Brand, publisher of counterculture bible The Whole Earth Catalog, and Larry Brilliant start an Bulletin Board System (BBS) to host what they term an online community. The Well attracts an eclectic mix of intellectuals, computer geeks, hippies, Grateful Dead fans, writers, entrepreneurs, and journalists. Journalists are given free memberships in the early days, leading to many articles about the community and helping it grow.

U.S. Internet protocols (TCP/IP) get a major boost when the National Science foundation forms the NSFNET, linking five supercomputer centers at Princeton University, Pittsburgh, University of California at San Diego, University of Illinois at Urbana-Champaign, and Cornell University. Soon, over a dozen regional and educational networks will be added, including BITNET, CSNET, and a dozen or so others. Parts of the original ARPAnet had been reassigned to NSFNET, while others had gone to the military network, MILNET. The NSFNET is a major factor in helping Internet protocols win out over rival protocols like OSI, SNA, and DECNET.

Digital mobile networks had been pioneered by ARPA from the early 1970s for military use, but early cell phone networks for consumers are analog. They use traditional telephone circuit-switching, where there is a connection (circuit) between caller and recipient for the duration of the call. The connection seamlessly switches from cell to cell as the phone moves. In 1987, the European Community formally agrees on the GSM (Spécial Mobile Group, or GSM in French) standard for digital mobile telephony, including text messaging (SMS). Development work had begun five years earlier with major input from Ericsson, Deutsche Telekom, France Telecom and a number of others. GSM will reach customers in the early 1990s, in some markets at around the same time as competing standard CDMA from Qualcomm in the United States.

Clicking on hyperlinks is what lets us “surf” the Web instead of plodding through it. Yet after its initial invention in the 1960s, computer hypertext had gone underground for 20 years. It got so obscure that the main father of the Web, Tim Berners-Lee, may have unknowingly re-invented it in 1980. Hypertext’s original inventors – and some true believers – had kept developing new applications, but mostly in academic environments or for specialized clients like the military. Without being named as such, hyperlinks had also been used in some online help systems and CD-ROMs.

By the late 1980s, a minor resurgence of interest leads to commercial hypertext programs like Owl – and then Apple’s well-hyped Hypercard. Soon, people are producing their own “stacks” of linked cards on every topic, as well as writing non-sequential hyper-novels. But HyperCard is standalone; you can only click through to other cards on the same computer.

23 year old Robert T. Morris, the son of a computer security expert for the National Security Agency, sends a nondestructive worm through the Internet causing major problems for days for about 6,000 of the 60,000 hosts linked to the network. The result is widespread outages. This is the first worm to have a major effect on real-world computer systems, and publicizes the importance of network security. Morris will be the first person convicted under the “Computer Fraud and Abuse Act.” He will apologize in 2008, saying he'd sought to estimate the Internet's size, not cause harm.

At its official 1983 launch, the Internet had been a modest experimental network of networks owned by the U.S. government. As late as 1989, even insiders are betting against it – OSI is the official favorite for the future of internetworking, or connecting networks together. But in the meantime the Internet has quietly grown to 100,000 host machines, each with multiple users. By 1992 the Internet will have emerged as the new global standard, linking a million computers. In hindsight, the Internet has several key advantages, from a growing community of enthusiasts churning out working software and hardware, to free distribution with the UNIX operating system, to being built in to common hardware like Cisco routers.

But the decisive factor? Probably money—especially U.S. government support from the National Science Foundation’s NSFNET and other sources. At the instigation of computer pioneers, Senator Al Gore begins working in 1987 on what will become his High Performance Computing and Communication Act. When it is funded in 1991, the Act creates the National Information Infrastructure, which promotes and funds over $600 million worth of various networking initiatives. Gore famously calls it the “information superhighway.”

The Internet connects over a million people by the end of the 1980s and is growing fast. But because it is a closed, non-commercial network used mostly by geeks, it lacks online systems to help ordinary people navigate it. None of the companies making slick, easy-to-use online systems like Minitel in France, CompuServe, AOL, etc. want to invest in porting them to an academic network. In any case they have their own networks.

This vacuum at the top of the Internet creates an opportunity for small players to try and create or adapt their own online systems. Usenet is the first; though mostly for geeks its discussion groups are quite popular and it gets ported to run over the Internet by 1986. Others range from low-key commercial ventures like WAIS and Hyper-G to student projects like Viola, Lynx, and Gopher. Several use clickable hypertext links – including one small experiment ambitiously called the “WorldWideWeb.”

At the world’s biggest physics laboratory, CERN in Switzerland, English programmer and physicist Tim Berners-Lee submits two proposals for what will become the Web, starting in March of 1989. Neither is approved. He proceeds anyway, with only unofficial support from his boss and his coworker Robert Cailliau. By Christmas of 1990 he has prototyped “WorldWideWeb” (as he writes it) in just three months on an advanced NeXT computer. It features a server, HTML, URLs, and the first browser. That browser also functions as an editor—like a word processor connected to the Internet – which reflects his original vision that the Web also incorporate authoring and personal organization tools. The idea is that a Web of useful links will grow and deepen as people create them in the course of their daily lives. The Web had been partly inspired by his earlier Enquire program, which had combined networked hypertext with ideas that would later evolve into the Semantic Web.

After the National Science Foundation (NSF) changes its policy, the Internet is for the first time a publicly accessible network with no commercial restrictions. This removes the last major remaining advantage for competing networking and internetworking standards, from OSI to SNA to CompuServe’s own international network. Four years later the NSF will turn over the Internet’s backbone (main high speed lines and nodes) completely to private industry.

Tim Berners-Lee’s 1990 GUI browser-editor runs only on rare NeXT computers. CERN refuses to fund other versions for common platforms. So the Web team writes a simple text-only browser for quick distribution, and then begs volunteers to write or adapt the needed GUI browsers for PCs, Macs, and UNIX machines. The team also provides code to start with; the WWW Common Library is essentially a build-your-own-browser toolkit written by Tim Berners-Lee and technical research assistant Jean-François Groff.

Eight volunteers respond, resulting in UNIX, Mac, and PC browsers. Viola and Midas are initially the most popular, eclipsed later by Mosaic. All of them leave out editing features, which are trickier to implement on machines other than the NeXT. Berners-Lee never regains control of his creation.

Gopher, which organizes content in folders rather than clickable links, grows faster than the Web in the early ‘90s and is its most direct Internet competitor. Educational institutions embrace Gopher, as do the U.S. Congress. Developed by Mark McCahill, Paul Lindner and Farhad Anklesaria from a Campus-Wide Information Service, Gopher is named both for the University of Minnosota mascot, and after “go for” meaning fetch. By 1993, the gopher developers are planning to add hyperlinks and even virtual reality features.

The Web pulls ahead partly by incorporating the ability to read Gopher pages; this is the same absorption strategy it had employed previously when it added support for WAIS and others. Two other rival standards, Lynx and Viola, have conveniently converted themselves into Web browsers. But with Gopher, the Web also gets a major lucky break: the University of Minnesota begins charging for Gopher server licenses in 1993, literally the same spring the Web becomes officially public domain – and free.

Mosaic, the first browser supported by a major institution, starts the Web on the road from research project to blockbuster success. Written by brilliant student Marc Andreessen and UNIX expert Eric Bina at the National Center for Supercomputing Applications, Mosaic was modeled on the Viola and Midas browsers and also used the CERN code library. But NCSA quickly assigned teams to write UNIX, Mac, and PC versions, as well as servers. Unlike other browsers it was reliable and could be installed by amateurs. Along with other browsers around this time Mosaic added graphics within Web pages instead of in separate windows. Mosaic spread like wildfire.

Business people are wary. How can you make money on the Web and the Internet? They are both open standards; you can't charge by the minute as online systems have done since their start in the 1960s.

In 1993, O’Reilly’s pioneering Global Network Navigator Web portal is running online ads. In 1994, Enterprise Integration Technologies (EIT) founds the CommerceNet consortium to encourage Web commerce, and demonstrates secure credit-card transactions that same year.

The first businesses to earn substantial profits on the Web are pornography and gambling sites, by 1995. But it is Netscape’s spectacular IPO, and the success of online shopping sites like Amazon and eBay, that finally convinces mainstream business to follow the pioneers into Web commerce.

When main Web inventor Tim Berners-Lee forms the World Wide Web Consortium (W3C) in 1994, the European headquarters are slated for the Web’s birthplace, CERN in Switzerland, with U.S. headquarters at MIT in Boston. But then CERN changes its plans and the core team of Web developers gets split among several French research sites. Also in 1994, Vice-President Al Gore supports a prominent White House Web site, as well as encouraging funding of W3C in the U.S.

Perhaps most important, Silicon Valley begins to invest in the commercial possibilities of the Web – including Java and the formation of Netscape. The momentum for Web development shifts further West, and never returns to Europe.

Browser War I had been more of a coup – when half the Mosaic team defected in early 1994 and formed Netscape under entrepreneur Jim Clark, Mosaic lasted less than a year. But when Microsoft licenses a version of Mosaic and rebrands it Internet Explorer, the fight is on. In the mid to late 1990s Netscape revolutionizes the business model for the Web, and helps it spread to ordinary people as well as businesses.

But Microsoft gives away Explorer free with every copy of Windows 95 and beyond, and by the end of the 1990s Netscape is failing. As a last-ditch strategy the code for Netscape's Navigator browser gets converted to open source, and becomes the basis of the Mozilla Foundation and its Firefox browser today.

Most of the big “walled gardens” — CompuServe, AOL, Minitel in France—resist the Web and Internet. By the mid 1990s they are either fading out or on their way to becoming Web portals.

Microsoft Network (MSN) is the one that might have mounted a serious challenge. The tens of millions of copies of Windows 95 come ready to connect to this private network, which has proprietary protocols; it could have become the biggest online service in the world nearly overnight.

But by 1995 the Web is growing quickly, and Microsoft CEO Bill Gates decides it is better to fight within the Web than to fight the Web itself. In a single memo, he turns company strategy completely around to focus on the Web in nearly every product. MSN becomes a Web portal.

At the end of 1996, the 36 million Web users surpass the 30 million or so on France’s Minitel, until now the most popular online system. By decade’s end, the Web will hit 360 million. By 2010, two billion.

Japanese mobile phone operator NTT DoCoMo creates the i-mode networking standard for mobile data in 1999. By 2002, over 34 million subscribers are using it on their phones for web access, e-mail, mobile payments, streaming video, and many other features that the rest of the world won't see for nearly another decade. The i-mode protocols, a simplified version of the standard HTML web language, are designed to work well with devices having small screens, limited buttons, and no keyboard. Related systems like WAP (Wireless Access Protocol) have fewer customers, but all of them contribute to bringing mobile browsing to a mass market.

In 1999, the growing IEEE 802.11b short-range radio networking standard is rebranded “Wi-Fi” by the Wi-Fi Alliance. This is the same year Apple releases its "Airport" Wi-Fi router and builds Wi-Fi connectivity into new Macs. These and other consumer products help popularize cable-free connections at work, in cafes, and at home.

As users flock to the Web, the opportunities seem boundless. Nearly everything you could do with previous networks is ported to the Web, and every business sector, community, religion, and subculture stakes out a place online. Initial skepticism gives way to experimentation, and then mounting excitement as people begin to believe that the old laws of business don't apply to this new medium. Nobody wants to be left behind, fueling a frenzy of business ventures—many built on shaky foundations.

In early 2000, business fundamentals reassert themselves. In one year, technology stocks lose about 60% of their value. The boom and bust have their greatest effect in the San Francisco Bay Area, home of Silicon Valley as well as many previous booms from the Gold Rush on.

In 2004, Google is the first major Web company to float a publicly traded stock since the go-go days of the dot-com boom. This is a direct result of Google solving the eternal problem plaguing all previous search engines – how to profit from search. The secret turns out to be a discreet form of advertising, based on auctioning off keywords to appear as "sponsored results" within a search results page. Many people take Google's Initial Public Offering (IPO) as a sign that the Web is not only back from its deep trough after the crash but entering a new period of expansion, and many other IPOs follow Beneath it all, of course, the Web continues to steadily grow as it has since the early 1990s.

The original Web concept, and many pre-Web systems, had depended heavily on user contributions. Yet many 1990s Web sites had been more like traditional TV or radio broadcasting, with providers feeding content to passive surfers. Partly this had been because dominant Web browsers lacked editing capability. From the early 2000s a number of sites begin helping users generate and shape content: wikis, blogs, social networking sites, and more. Photo and video sharing sites take advantage of the spread of faster Internet connections to let users both upload and browse those media.

O’Reilly and Associates popularizes the name “Web 2.0” with their 2004 conference of that name. Most browsers still don't support Web page editing, but Web 2.0 sites find various workarounds – from wiki and blogging software to commenting features – to give users a voice.

In the 1960s when computers were extremely expensive, a number of companies had offered what were called computer utilities. They would run your programs and store your data on their computer, which you would access with a terminal. As time went on cheaper computers had made it more economical for companies and eventually individuals to maintain their own workstations and PCs. But in the Web era, the economies of scale that evolved from large commercial Web servers had begun to tip the balance back the other way.

Starting in the mid 2000s the computer utility model starts to became fashionable again under the name “The Cloud,” and is once again a major trend in both networking and computing. Amazon's 2006 Elastic Compute Cloud helps popularize the idea. Today, cloud-based companies offer nearly any software or service – including data storage – that could be done on a personal computer or on larger machines run by a company’s IT department.

By the late 2000s, 3G networks for higher speed mobile data had been spreading fast. The iPhone’s phenomenal popularity creates a new computing platform that brings mobile Web browsing to a large audience. Google’s Android mobile platform soon makes that audience even larger. The App store model used by the iPhone and then Android is based on Apple’s earlier success with iTunes. But because proprietary apps run directly over the Internet, they are not part of the public Web – and present a risk of fragementing it as a standard.