The World Wide Web as an Engineering Paradigm 
Matthew Lee Andrew Montgomery 
Steven Shapiro Veeral Shah 
Qian Wang 

December 15, 2000 
6.933/ STS. 420 – The Structure of Engineering Revolutions Professor David Mindell and Professor George Pratt 1 
1 Page 2 3 
THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
2 
Introduction 
"You see things and say 'Why? ' 
but I dream things that never were, and I say 'Why not? '" 


George Bernard Shaw (1856 – 1950) 


The World Wide Web is a compelling example of an engineering paradigm. The 
adoption of the Hypertext Markup Language (HTML) and the Hypertext Transfer 
Protocol (HTTP) as standards represent the acceptance of this new paradigm— a 
revolution in information sharing. This work details the technical evolution of HTML 
and HTTP, from their inception to their acceptance as the technological foundation of the 
World Wide Web.

This project history first equates the Web's development an engineering paradigm, 
with a discussion of how it departs from the Kuhnian model of scientific paradigms. 
Next, a history of the important events preceding the development of the Web sets the 
stage for an understanding of the background in which its creation took place. This will 
help to illuminate the presence of particular anomalies. We will then examine the period 
of crisis that would ensue— the question of how to share information across multiple 
platforms. In the presence of this crisis, competing articulations and solutions arose— 
such as Gopher, WAIS, and Guide. Finally, this history delineates the rapid acceptance 
of the World Wide Web— the adoption of a new paradigm.

By no means did the World Wide Web signal the advent of online 
communications. The core technologies that the Web is based upon, such as TCP/ IP and 
hypertext, existed for years and in some cases decades before the Web was invented. 
Files were transferred with FTP, network news was read through NNTP, and network 
packets circled the globe through the Internet. Similar technologies to the Web existed as 
well, including Gopher, a system developed at the University of Minnesota. 

Despite all of these perfectly functional technologies, the Internet was fraught 
with incompatibilities, as data was stored in a myriad of often-incompatible formats and 
the usage of protocols was inconsistent across different platforms. For example, the 
simple act of retrieving a file could prove troublesome. First, one would have to locate a 
remote server that the file would be hosted on, through a primitive search service such as 
Archie or by being told about a server from a colleague. Then the file would have to be 
retrieved via FTP, which involves a series of directory navigation commands and some 
protocol negotiation as well. Finally, the file would have to be viewed locally, and the 
viewer might be incompatible or nonexistent for a particular format, especially when the 
file was created on another platform. Tim Berners-Lee invented the Web in part to 
resolve these incompatibilities, relying on a small set of simple protocols to share data in 
a platform-independent manner. This is the story of the Web. 


Scientific Revolutions 
Thomas Kuhn, in his seminal work "The Structure of Scientific Revolutions," outlined 
the development of a paradigm:

1) Paradigm / Pre -Paradigmatic state: At this time, there may or may not be a 
conceptual framework by which people assessed the day's ideas. This framework 
represents the "agreed upon world view" at the time. Within this state, the 
progress that takes place is "normal science:" these developments are incremental, 
and do not challenge the prevailing paradigms. 
2) Anomaly: An incompatibility with the existing paradigm arises. Evidence or 
observations appear to contradict the currently understood framework.
3) Crisis State: This stage is the reaction to the anomaly. The period can be 
characterized by confusion, and the existing paradigm is challenged. Competing 
articulations emerge, to try to explain the newly found evidence. 
4) Resolution: The resolution of the crisis may take one of three forms: 
a. Postponement – The anomaly is put aside because it is not yet well 
understood. 
b. Integration – The new phenomenon is sometimes found to be compatible 
with the previously existing paradigm, and the two are integrated into a 
fuller, more complete paradigm. 
c. Paradigm Shift – This is the most momentous of the three types of 
reactions. Out of the competing solutions emerges an undisputed new 
consensus. It is this consensus that defines the new conceptual framework 
as the prevailing paradigm.

Engineering Revolutions 
The characteristics of engineering revolutions differ along some important metrics from 
the Kuhnian model, which is more suitable for describe scientific change. Noted here are 
several key differences.

Technology Driven: The primary difference between scientific and technological 
revolutions is, quite simply, the artifacts at hand. In scientific revolutions, what 
previously exists is a framework upon which to base thought processes and research. In 
contrast, technological revolutions involve technological artifacts, with a newer 'artifact' 
dominating. In these cases, both artifacts work properly, and provide some benefit to 
their users, but the mere existence (and growing usage) of the newer technology instead 
drives the revolution.

Not Fundamental or Global: Technological revolutions are characterized by a 'better 
way of doing things. ' Unlike the acceptance of a broad scientific paradigm, the 
acceptance of a new engineering paradigm does not have fundamental implications on the 
whole world. 

Consensus is Not Necessarily Universal: Often times a new engineering paradigm is not 
universally accepted. It is only important that the new framework dominates its field. 

The rest of this paper explores the advent of the World Wide Web, providing a chronicle 
of its development within the defined technological revolution framework. 

The Existing Information Sharing Paradigm 

We have just introduced our concept of the structure of engineering revolutions. 
Any such revolution must begin with either an existing paradigm or a pre-paradigmatic 
state that resembles the crisis state we have discussed. The engineering revolution that is 
the focus of this paper is that of the World Wide Web and its revolution of the existing 
information-sharing paradigm. Before discussing the Web itself, however, we must first 
examine the paradigm immediately preceding the Web, the new technologies which arose 
that presented an anomaly in the paradigm, the crisis state that was created by this 
anomaly and the competing solutions that arose to solve the crisis. We must understand 
what occurred before the Web in order to understand why the Web was developed and 
why it is the way to share information electronically today.

Computer systems prior to the 1960s consisted mostly of mainframes at large 
research and educational institutions and businesses. No one had conceived of a personal 
computer as we understand them today. The systems first used punch-cards as a method 
of input. Each system had a designated operator that would feed it the cards dictating 
operations that various users wanted performed. Only one set of cards could be inputed at 
a time and the system could only do one set of operations. But, the systems were very 
large and relatively fast and could handle much more work. Thus, timesharing of the 
systems began, in which a terminal with a wire connecting it to the mainframe could be 
placed on a users desk and punch cards were eliminated. Each system used its own 
operating system and protocols. There was no need for the systems to be compatible with 
each other, as there was no network to connect them. Data was thus confined to the local 
machines and information could only be shared within the institution which housed the 
mainframe. Vannevar Bush, director of the US Office of Scientific Research and 
Development under President Roosevelt, wrote an article in 1945 titled As We May Think 
published in Atlantic Monthly. These were the some of the first thoughts that were 
published thinking about the power of computers in information sharing. His "Memex" 
sounds a great deal like today's personal computers and the user's interaction with the 
system very much resembles that of hypertext, even though no one would "invent" either 
technology for decades.

This leads us to the three technologies that created an anomaly in the information 
sharing paradigm. Collaboration and hypertext systems would give people a way to 
organize and share their information. Pervasive networking would allow people to 
connect their systems together over large distances. Personal computers would allow any 
person, at the office or at home, to share their information. Not until all three of these 
technologies were developed was there really an anomaly in the information sharing 
paradigm. We will now examine each of these technologies in more detail in order to 
understand this anomaly.

New Technologies Cause an Anomaly 

Collaboration and Hypertext Systems

If it was not for Vannevar Bush's provocative article, As We May Think, Doug 
Engelbart may not have invented the mouse or formed the idea of the first hypertext 
system. Engelbart first stumbled across the Vannevar Bush article and recalls that it had a 
profound effect on him. "I remember being thrilled," he says. "Just the whole concept of 
helping people work and think that way just excited me. I never have forgotten that." 1 He 
was inspired and began to work on his oN-Line System (NLS) in 1963 and first publicly 
demonstrated it in 1968 at the Fall Joint Computer Conference in San Francisco. The 
system really was revolutionary for its time; it incorporated a graphical display, video
conferencing, hyperlinks for organizing information, and the computer mouse (which he 
invented for the project) for navigation. See Figure 1, a screen shot of Engelbart's 
demonstration of NLS in 1968. NLS was a completely successful implementation of what 
Engelbart had imagined and people congratulated him for it. However, the same people 
just could not see the purpose of it given their view of the role of computers: 
large mainframes using punch cards. Engelbart was a man far ahead of his time. Decades 
later, when asked about the lack of acceptance for NLS, he says "You can't bring stuff
out against the prevailing paradigms." It would be more than twenty years before many
of Engelbart's ideas would be used to force a paradigm shift.

Other collaboration and hypertext systems were also in development during the 
1960s. In fact, Ted Nelson would coin the term "hypertext" during a presentation at the 
1965 Association for Computing Machinery conference in Pittsburgh. Ted Nelson would 
be one of the first insightful individuals to develop Doug Engelbart and Vannevar Bush's 
shared vision. Nelson first developed a system called Xanadu, which he described as a 
computerized version of Bush's Memex. Xanadu will be discussed in more detail 
shortly. Nelson went to work with Andy Van Damm and some of his students at Brown 
in 1967 to create the second hypertext system, the Hypertext Editing System (HES). This 
was not an entirely significant hypertext system, but more importantly Andy Van Damm 
and his students continued, after Nelson stopped collaborating in 1968, to work on a new 
system called the File Retrieval and Editing System (FRESS). Although FRESS had no 
networking (the first ARPANET node was not even online yet), it had two things that 
later hypertext systems would find valuable. First, it had bi-directional links that allowed 
for the user to control the link structure without modifying the original document content. 
The second was a form of metadata that allowed keyword searching of links. Other 
systems that were to follow, such as Microcosm and Intermedia, borrowed from these 
ideas separating the links from the content. FRESS proved to be a relative success in the 
classroom for which it was designed. However, FRESS and other similar systems would 
not solve the information sharing crisis. The nature of these systems was to separate the 
links from the original source documents and maintain them in a central database. This 
database would have to be continuously maintained in order to retain the consistency of 
the bi-directional links. This, in turn, would prevent FRESS and other similar systems 
from scaling to a large (especially global) scale. 

Ted Nelson's Xanadu project continued for over two decades after its original 
inception. It provided a framework for information sharing, culture and management. 
Xanadu's biggest difference from other systems of the time was its micropayment 
scheme. Xanadu's creator was not a member of the computer science community [say 
what he was here]. This may have been what caused Nelson to heavily focus on giving 
credit to the author of every bit of information. It is suspected that Xanadu never really 
became a complete system because of its visionary nature. A product must be concrete. 
Xanadu presented itself almost as an entirely different way of life. Nelson's writing on 
Xanadu, "Literary Machines," made references not only to information management, but 
what people would think and how they would dress. Xanadu was just too far reaching to 
be seriously considered for a viable solution to the engineering crisis at hand. 

Pervasive Networking

On October 4 th 1957, the Soviet Union made history when they launched Sputnik 
I into orbit. President Eisenhower declared that the United States would never again be 
caught off guard by the Soviet Union and went to Congress on January 7 th , 1958 to 
request funding for a new research and development organization. Congress quickly 
approved funding for the new Advanced Research Projects Agency (ARPA) and, by 
1962, ARPA's long-term goals had been refined. The two major parts of ARPA involved 
ballistic missile defense and nuclear test detection. However, there was a small office, 
using less than 10 per cent of ARPA's budget, called the Information Processing 
Techniques Office (IPTO) that by 1965 was immersed in research for advanced 
computing. Computer time-sharing among many different systems across the country was 
the solution to the lack of funding the IPTO had and the multitude of requests from 
computer science departments who each wanted to build very powerful computers. By 
sharing the computer resources, more could be done. Thus, the IPTO began work on a 
nationwide network to connect the various systems. By late 1966, plans for the ARPA 
Computer Network, or ARPANET, had begun.

It was not immediately clear how the network should be built. Initial ideas from 
the project's director, Larry Roberts, focused on just having each computer call up all the 
other computers using the existing phone lines to exchange information. This would force 
each computer to understand the operating system of all the others. There are obviously 
many problems with this approach, as the network could not scale very well or be 
flexible. Wesley Clark, an ex-colleague of Roberts, came up with the idea of having 
small computers on the network to route data around, or sub-nets. This would mean the 
large computers on the network would only have to know one protocol or language to 
talk to each other – that of the small routing computers. The small computers were named 
Interface Message Processors (IMP) and their use was essential to the success of 
ARPANET. Another important part of the design or ARPANET was that it used packet-switching. 
This new idea, developed independently by both Paul Baran and Donald 
Davies, cuts messages up into small chunks of the same size for transmission across the 
network. It allowed a network to exist that would be able to survive machine failure on 
one part of the network; the packets could just be rerouted through different parts of the 
network.

UCLA was the chosen location for the first IMP and the second was put into place 
at the Stanford Research Institute (SRI) in October 1969 and the first packet-switching 
computer network was established. ARPANET grew quickly, adding about one host per month.
In 1972, the first email was sent across ARPANET and the network was publicly demonstrated
for the first time at the first International Conference on Computer Communication (ICCC)
in Washington, DC. See Figure 2 for a picture of ARPANET in 1977.

The first protocols on the ARPANET were hastily thrown together; they were 
telnet and FTP. Telnet allowed someone sitting at a terminal to login to a computer at 
another site. FTP, File Transfer Protocol, allows files to be transferred from one computer 
to another. These two protocols were a good start, but hardly sufficient for the network. 
In June 1973, Vint Cerf, Bob Kahn, Bob Metcalfe and Gιrard Le Lann started to work on 
the design of a host-to-host protocol for inter-networking. They called it the Transfer 
Control Protocol (TCP). TCP proved to be a robust host-to-host protocol that allows 
reliable communication over unreliable networks. In 1978, TCP was combined with 
another protocol, Internet Protocol (IP), to form TCP/ IP. IP controlled the routing of 
packets around the network. Thus, with the introduction of the Domain Name System 
(DNS) by Paul Mockapetris and the adoption of TCP/ IP by ARPANET in 1983, the 
Internet as it is known today was born.
 
This pervasive network, the Internet, allowed any network of computers to be 
connected to any other network of computers. People could now share information 
electronically, using a universally agreed-upon protocol. But, in the early 1980s, 
computers were still primarily at large institutions and there was a relatively small group 
of users. Not until the rise of personal computers would there by a true crisis in the 
paradigm of information sharing.

Personal Computers

Computers used in the existing information sharing paradigm were massive in 
both size and price. Only large institutions could afford and support them. But, the 
invention of the transistor in 1947, the integrated circuit in 1958, the microprocessor 
in 1969, and magnetic memory in 1972, set the stage for computers to become smaller and 
cheaper. They would become a part of most people's daily lives– a PC would soon sit on 
their desk at work and at home.

The Xerox Palo Alto Research Center (PARC) began work on its Alto computer 
in 1973. It was the first computer designed specifically for use by non-computer 
scientists. It actually was the first system to use many of Doug Engelbart's ideas, most 
notably the mouse. It had a graphical user interface (GUI) with windows, menus and 
icons. It looked very much like today's PCs. The machine that evolved from the Alto, the 
Xerox Star, was the first system that Xerox tried to sell, but very few were actually sold 
because of its $18,000 price tag.

In January 1975, an almost completely different system was released. The Altair 
8800, released by Micro Instrumentation Telemetry Systems (MITS), sold for $397. 
Users had to assemble the system themselves and the system was essentially a series of 
switches and lights. It was completely unclear how it even worked. But, many were sold 
to those who were curious about computers and had never had the chance to be exposed 
to them. It inspired them to be creative and innovate. Bill Gates and Paul Allen, for 
example, saw the Altair and formed Microsoft to sell software that allowed BASIC 
programming on the system. Steve Wozniak saw the Altair and was determined to build 
something better; the cost of microprocessors and memory had fallen dramatically. 
Wozniak joined forces with Steve Jobs to form Apple Computer and they released the 
Apple II at the West Coast Computer Faire in May 1977, selling it for a more reasonable 
$1300. Many Apple II's sold quickly and many companies started to develop software for 
it.

The early 1980s saw the real birth of the personal computer as it is thought of today. 
In 1981, IBM launched its PC, selling for $2,800. Businesses loved the PC. IBM's 
decision to allow clones ensured rapid growth of the technology. Microsoft wrote the 
operating system, MS-DOS, and would later write the Windows operating system. During 
this same timeframe, Apple was releasing its next computer. In 1984, Apple released its 
first Macintosh computer, selling it for $2,495. It was the first computer that brought
the technology that Xerox had developed based on Engelbart's ideas to the masses. The GUI 
stunned people and the Macintosh quickly developed a strong following amongst desktop 
publishers, something that it still maintains today. The IBM PC and the Macintosh 
changed the face of personal computing forever. 

A Crisis Develops

The exponential growth of personal computers into the 1990s signaled the 
establishment of the third technology that was needed for an anomaly in the information 
sharing paradigm. Recall that an anomaly occurs in an engineering paradigm when an 
incompatibility with the existing paradigm arises. In this case, information that was 
previously used by few individuals and stored on isolated systems suddenly had the 
potential to be used by many people and on any system. The three technologies – 
collaboration/ hypertext systems, pervasive networking, and personal computers created 
this anomaly. But, there was no single solution to solve the problem created by this 
anomaly. No system was commonly accepted that allowed people to collaborate across a 
network using their PC. Until such a solution was developed, the anomaly brought about 
a crisis in the information sharing paradigm. Competing articulations for solving the 
anomaly is characteristic of a crisis state and we will now discuss those competing 
solutions.
 
Any solution to this crisis state would have to have three characteristics for it to 
be successful. First, it must possess the ability to discover the information that is desired. 
The Internet allowed all the networks in the world to be connected, but provided no real 
means for finding desired information scattered across the millions of machines that were 
connected. Secondly, any solution would have to be able to retrieve the information 
regardless of the operating system and software being used. A standard protocol 
implemented on all platforms would be essential for success. Finally, the new system 
would have to be able to display the information uniformly once it was received, again 
regardless of the platform being used. Several solutions were developed that shared all of 
three of these characteristics.

Competing Solutions

Some would argue that creativity is at its peak during an engineering crisis state. 
In the case of the crisis being described in this paper, we will see that this is the case. 
There were many competing articulations that attempted to solve the discovery, display, 
and retrieval crisis in information sharing described above. However, these solutions 
solved the problem in a variety of creative and differing manners. There were more 
competing articulations than can adequately researched for the scope of this paper, but 
several, such as WAIS, Archie, Guide, and Gopher, will be discussed in the following 
sections. Two solutions are particularly relevant to development of the World Wide Web 
project as well as the question of 'what could have been if only... ', Office Workstation 
Limited's Guide and Mark McCahill's Gopher.

Wide Area Information Servers

Wide Area Information Servers (WAIS) could be construed as the opposite of 
Xanadu. WAIS was concrete, concise and focused with a commercial marketplace. It 
was the first successful information management system that was a commercial venture 
from the start. Brewster Kahle, then working for the supercomputing company Thinking 
Machines, was the head of the project. WAIS was similar to a modern day search engine. 
It's key benefit was its ease of use and relatively intuitive interface. Figure 5 below is 
one example screenshot of the first WAIS client from a paper written in part by Brewster 
Kahle published in Electronic Networking the journal. 3 Users could submit their query in 
plain English. Since files were described by keywords, media such as video or sound 
files could be searched as well as text. WAIS proved very successful and soon spun off 
into WAIS Inc. which was sold in 1995 to America Online for $15 million. WAIS Inc. 
was one of the first Internet companies and proved to many skeptics that an Internet 
company was possible. WAIS did not become the new paradigm because it was missing 
one key ingredient, a humanitarian vision. WAIS Inc. and Brewster Kahle's vision were 
not compatible with a free world-wide solution. The company was merely creating a 
product for profit. 

Archie, Prospero and FTP

Telnet and FTP were the first protocols on ARPANET and Neither protocol's 
clients were robust or easy to use. This is where Archie came into play as a possible 
solution to the ongoing crisis. In order for users to find programs on the Internet in 1983, 
they would be forced to conduct manual searches of known FTP servers. Even then, the 
filenames of these programs may not yield useful descriptions. Alan Emtage, a student at 
McGill University in Montreal, Canada, sought to alleviate some of the mundane 
searching with Archie in 1983. Archie is a few shell scripts that automate this search. As 
Archie quickly became popular, Clifford Neuman, a Ph. D. student at the University of 
Washington, saw his chance to improve upon Archie by integrating it with his distributed 
filesystem, Prospero. Prospero was used to organize Archie's files as well as expedite the 
discovery process. This system quickly became very popular because of the growing 
urgency for a new paradigm. The Prospero-Archie-FTP system adequately solved 
discovery, display and retrieval problems, but it did not become the new paradigm. 
"Only hardcore users knew about FTP." 4 This system lacked the necessary easy of use 
for the general computer illiterate population. In addition, although it was a breakthrough 
for its time, the system left much in the way of improvement especially with respect to its 
display of information. Long filename lists with descriptions are hardly comparable to 
the media rich interface of the Web as we know it today.

Real Solutions

Two contemporaries of the web deserve a more detailed look in this paper 
because of how close they actually became to becoming the new paradigm. Interviews 
with the systems parents and corroborative documents shed light onto this rarely 
documented turn of the Web. Tim Berners-Lee, the undisputed father of the World Wide 
Web, approached both Office Workstations Limited (OWL) and the Gopher Team to try 
to convince them to adapt their systems to form to his humanitarian vision of a single 
global information space.

Guide

Guide was first developed as an academic project by Peter Brown in 1983 at the 
University of Kent in the United Kingdom. This academic version of the hypertext 
system Guide worked a little differently than its later commercialized counterpart:
 
The way it works is that when you hit a button it replaces that button in-line so there is a sort
of metaphor in which you have an expanding document; you hit a button and that button expands, in-line, so it remains in the context where it was in the whole document. If you undo it, it goes 
back to the previous [context].
 
Ian Ritchie, the founder of Office Workstations Limited (OWL), purchased Guide from 
the Peter Brown and the university with a sole license to the system.
 
OWL released their version of Guide for the Macintosh in August 1986, winning 
the honor of becoming the first commercially successful hypertext system. Almost 
exactly a year later, OWL released an identical version of Guide for the PC. However, 
Apple Computer, Inc. forced OWL to concentrate on their newly released PC version of 
Guide. In the middle of 1987, Apple released their hypertext system, Hypercard, free 
with every Macintosh. Ian Ritchie notes in reference to OWL's shift to the PC market, 
"You can compete with other things, but you can't really compete with free." 6 This idea 
holds true in many other cases described in this paper.

OWL's implementation of Guide was extremely robust and much more advanced 
than Berners-Lee's World Wide Web. Guide used a markup language called the 
Hypertext Markup Language (HML) not to be confused with Berners-Lee's HTML. Ian 
Ritchie describes HML:

That was a version of SGML with added tags to indicate things like a button or a button source or a point, headers and body text, so forth.. And by marking up your documentation in that way you could automatically generate a hypertext display. And you could go from document to document. And if your documents were on a file server you could go across a network to them.. We did that… I mean the interesting thing now is that everybody is really adopting XML. But XML is.. essentially contains the kind of structures that we and other people had done in the late 80's with our markup languages. I mean our markup language, old HML, was 
more than just a hypertext markup language. It actually was a content structure language as well. It described for you, what was a heading.. what was a subheading.. what was an author's name.. what was an abstract.. what was a caption.. Yeah, that's the kind of thing that XML does. In fact there's evidence of that in a variant we were doing for <sounded like Verfe Ville Industry>. Cause we put the first thing that Ford Motor adopted for all of their technical publishing in North America. And in fact, later, they also were 
adopted by General Motors. So we had both General Motors and Ford. And there we just adopted their automobile markup language. They had a language they'd determined for marking up documents for automobile information. 

And that's.. All that stuff was very similar to what people today call XML.

OWL even produced Guidex (originally IDEX), which was a solution for a particular client which was a networked version of Guide. Ian Ritchie even described a similar version of Guide for a client that worked over the Internet! It is mind boggling to look at this little-known hypertext system that was much more advanced than the famous World Wide Web before the Web was really started.

Tim Berners-Lee was looking for a jump start on his Web project in an already 
existent system. He finally saw one that peaked his interest when Guide was 
demonstrated at the first European Hypertext Conference at Versaille in Paris in 1990. 
Ian Ritchie remembers it clearly:
 
What Tim wanted was a graphical user interface to make that more user friendly. And we had a chat about it.. Actually, I mean this was 


hardly a summit meeting.. He and I met in the bar at the hypertext conference in Versaille and he mentioned this project that he was 
working on called the World Wide Web. I remember thinking it was a rather grand name. 


Yeah, it did seem like a bit of a grand name at the time. But he said that they were working on that stuff and they were taking off 
great well.. and he wondered if we might want to, you know, share in some of this work. He was particularly interested in the work 
we'd been doing. He'd read about the work we'd been doing in GuideX because he heard we had a markup language. So I agree 
that I would send him details of that. And we sent him various specs and so forth of the markup language. We had a sort of exchange of 
correspondence. 8 
However, Guide and Guidex were not public domain, and OWL was a for-profit 
company. CERN was in no position to purchase external engineering as they were 
devoting most of their budget to particle accelerator projects and other expensive physics 
equipment. If CERN had purchased Guide from OWL it quite possibly could have 
become what we think of as the World Wide Web today. However, Guide was much 
more advanced than the Web, so some of the well established problems of the Web today 
(such as dead links) may have already been solved. 


Gopher 
In the late 1980's all major universities had embraced the Internet and begun to 


8 Ian Ritchie phone interview. 17 
17 Page 18 19 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
18 
digitize various content. These institutions were plagued by the same discovery, display 
and retrieval crisis as the rest of the networked community. In the early 1990's, Campus 
Wide Information Systems (CWIS) seemed to be the obvious solution. These systems 
were very different, but they all had a common goal: To organize the vast microcosm of 
the university's information space. 
The University of Minnesota was no exception. A lot of people had an interest in 
participating in the university's effort, so a committee was formed to design their ideal 
system. Mark McCahill, an employee at the university's computing center, was 
eventually given the task of managing the project. In short, McCahill and his lead 
developer, Farhad Anklesaria, felt the messy decree delivered by the committee was 
unacceptable and followed their own design. Their CWIS was named Gopher, after the 
university's mascot, and McCahill and Anklesaria went to the committee in 1991. 
Gopher met with enormous resistance from the University committee because it did not 
follow their decree. The University of Minnesota rejected Gopher. However, the system 
was still released to the public domain in June of 1991. Gopher quickly became the 
Web's biggest contemporary competitor and was the first solution to the impending crisis 
that really opened the internet to anyone.. It was simple, intuitive and well-structured. A 
hierarchy of menus whose items can be submenus or an object (like a document, picture 
or any other object). "The information for any given item is enough to identify the server 
it's on" (p. 138 h. w. b.) This feature was integral to the success of Gopher because it 
meant the information did not have to be stored on the computer you were using. One 
thing had to be added before Gopher's preliminary release, full-text database searching. 
Tim Berners-Lee also recognized the value of Gopher as a world-wide system and 
approached the Gopher team at the 1991 Internet and Engineering Task Force (IETF) 
meeting.. 
Tim was all excited and said 'Look at this stuff I came up with! We should work together because we're doing a lot of the 


same things. ' But when I got back and looked at it some more, I said, 'Shit, I don't work this way! ' 9 


Mark McCahill was quick to reject Tim Berners-Lee's ideas for one very good 
9 Mark McCahill phone interview. 18 
18 Page 19 20 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
19 
reason: the Web project lacked structure. In fact, from McCahill's description, the 
Gopher team would have rejected any idea to migrate to a hypertext system where objects 
can be randomly associated. Gopher's biggest strength was its structure and ability to 
organize information for linear thinkers. By 1993, Gopher had exhibited exponential 
growth in its user base. "In April of that year [1993], there were just over 460 registered 
gopher servers; by December there were over 4800". 10 Soon, the University of 
Minnesota's computing center was forced to accept Gopher as it took off around the 
country's Universities. This was the system that could have been the new paradigm. It 
had a jump start on the WWW with a substantial user base. However, the Gopher team 
would soon make a big mistake. 
Companies were going to the University and wanted to use Gopher for their 
systems. So, the university decided to start charging everyone except for non-profit and 
academic institutions. Even though the university was not charging everyone, they still 
crossed the line between General Public Domain (GPD) and a licensed product. "It 
pissed a lot of people off," remembers McCahill. "It was a funny time. There was a 
recession on and the school just didn't want to give stuff away for free." 11 The gopher 
team got flamed in a monumental way. In 1994, Gopher lost its ground relatively quickly 
and eventually got consumed by the WWW. 


Tim Berners-Lee's Vision 
Tim Berners-Lee was the son of two mathematicians who 
programmed some of the world's first computers. As Berners-Lee 
notes in his recent book, "They were full of excitement over the 
idea that, in principle, a person could program a computer to do 
most anything." He mentioned that he was watching his father 
write a speech for Basil de Ferranti (of Ferranti Computers, a seller 
of the early Mark I machine). His father was researching on the 
brain, thinking about the random associations that the brain makes and exploring the 


10 How the Web was Born, p. 140 
11 Mark McCahill phone interview. 


Tim Berners-Lee 19 
19 Page 20 21 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
20 
possibility of computers behaving similarly. The idea struck Berners-Lee and he 
continued to think about that topic throughout his life. 12 
Berners-Lee continued in his education at Oxford, graduating in 1976 with a 
bachelor's in physics. After a few years at Plessey Telecommunications and DG Nash, 
Tim took a consulting job in 1980 with CERN, the European particle physics laboratory 
in Geneva. Here, Berners-Lee wrote his first attempt at information management, a 
program he called Enquire. He just wrote it in his spare time for personal use, to 
remember the connections at CERN among the various computers, projects and people at 
the lab. Berners-Lee credits this as the time that a larger vision took root in his head. 
"Suppose all the information stored on computers everywhere were linked. 
Suppose I could program my computer to create a space in which anything could be 
linked to anything. All the bits of information in every computer at CERN, and on the 
planet, would be available to me and anyone else. There would be a single, global 
information space." 
After only six months at CERN, Berners-Lee returned to England to work with 
John Poole, a former colleague. Poole had formed a company based on dot-matrix 
printer control software that Poole and Berners-Lee had worked on two years earlier. 
Berners-Lee helped enhance the software and bring it to market, including developing a 
special markup language for document preparation. However, in the spring of 1983, 
Berners-Lee needed a change from living in Britain, and applied for a fellowship back at 
CERN. 13 
Berners-Lee returned to CERN in 1984. By this time, the growth of networking 
and various protocols had continued to the point of sustaining various computers at 
CERN. As Berners-Lee mentions in his book, "Machines from IBM, DEC, Control Data--
we had them all, as well as the new choice of PC or Mac in personal computers and 
different word processors." In this atmosphere, he wrote some RPC (remote procedure 
call) tools. 
Given Berners-Lee's experience with designing Enquire and Tangle, and after 
reflecting upon the need for a new documentation system at CERN, he was able to 


12 Weaving the Web, pgs. 3-4 
13 Weaving the Web, pgs. 11-12 20 
20 Page 21 22 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
21 
formulate a vision of a new system that would become the World Wide Web. The grand 
design for the system was for it to be a universal information space -a system that could 
store anything and the links between pieces of information. 14 Berners-Lee decided on a 
web-like structure using hypertext, in which nodes of information were connected by 
links, which could connect any two pieces of information together to describe a 
relationship between those pieces of information. By completely abandoning any notion 
of constrained structure, and by using simple protocols and formats, Berners-Lee created 
a system that could be acceptable to all -"a system with as few rules as possible." 15 
To represent the nodes in this web of information, Berners-Lee chose a simple 
hypertext document format, represented by standard ASCII text documents. Special 
spans of text in each document, such as headings and links, were set off from the rest of 
the text using simple tags like 

<P ALIGN=center CLASS="style1" STYLE="line-height: 150%">Welcome</P>
<P ALIGN=center CLASS="style1" STYLE="line-height: 150%"><a href="http://www.larryblakeley.com/larryblakeley20070521.jpg">Larry Blakeley (May 21, 2007)</a></p>
<P ALIGN=center CLASS="style1" STYLE="line-height: 150%">http://www.larryblakeley.com/larryblakeley20070521.jpg </p>
<p align="center" class="style1" style="line-height: 150%">(Contact Info: larry@larryblakeley.com)</p>
<p align="left" class="style1" style="line-height: 150%">Important Note: You will need to click this icon to download the free <a href="http://www.celartem.com/en/download/djvu.asp"><img src="http://www.royblakeley.name/dl_djvu.jpg" width="102" height="28" border="0"></a> needed to view most of the images on this Web site - just a couple of clicks and you're "good to go." <p align="left" class="style1" style="line-height: 150%">I manage this Web site and the following Web sites: <a href="http://www.leslieadkins.name/index.htm">Leslie (Blakeley) Adkins</a> - my oldest daughter</p>
  <p align="left" class="style1" style="line-height: 150%"><a href="http://www.loriblakeley.name/index.htm">Lori Ann Blakeley</a> (June 20, 1985 - May 4, 2005) - my middle daughter</p>
  <p align="left" class="style1" style="line-height: 150%"><a href="http://www.evanblakeley.name/index.htm">Evan Blakeley</a>- my youngest child</p></ BODY>. The most important aspect of this 
representation of hypertext is that it is extremely platform-neutral, as plain text can be 
read by every computer platform. Berners-Lee realized this, and used the cross-platform 
nature of ASCII text to truly exploit the diversity of different computer systems and 
networks that could possibly use the Web. 16 
Following from the establishment of this arbitrary web structure for his universal 
information space, Berners-Lee made the key insight that since any node could be linked 
to another node, there existed an equivalency between all nodes from a systems point of 
view. A page buried deep inside a proprietary help system would be no different from a 
transcript typed into a flat text file --each node would need to have a unique address, 
allowing a viewer to bridge across information systems seamlessly within a single 
viewer. In addition to unifying different data formats, this common addressing scheme 
sought to unify various hardware platforms as well, since the address would be platform-neutral. 17 


CERN's organizational structure dictated another important aspect of this new 
system. Since CERN's scientists were constantly coming and going, and doing work 


14 Weaving the Web, pg. 16 
15 Weaving the Web, pg. 15 
16 Weaving the Web, pg. 15 
17 Weaving the Web, pg. 16 21 
21 Page 22 23 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
22 
from remote locations 18 , it was essential that the new information system be network-aware. 
This measure strengthened the platform-independence of the system, as common 
networking protocols are often used to connect disparate systems together. TCP/ IP was 
chosen because it was used already by the Unix world, and VAX machines could be 
patched to support TCP/ IP as well. 19 However, perhaps the most important and desirable 
feature of a network-distributed system is that new systems can come online without 
disrupting or modifying any existing systems, allowing the system as a whole to scale 
with the size of the network. 20 
Scalability was a key issue that Berners-Lee was aware of from the beginning. It 
proved to be the downfall of many more robust systems attempting to solve the 
information management crisis. He realized that any one central control point for any 
aspect of the system would limit its ability to scale up to thousands or millions of 
users/ servers. It could also limit access to certain groups of people. 21 A key aspect of 
decentralization was the realization that a central link database would fatally cripple the 
system's ability to scale. The benefits of a central link database is that there would never 
be a broken link; however, Berners-Lee decided that the ability to scale up was worth the 
occasional inconvenience of broken links. 22 
Collaboration, and more generally, online editing of hypertext documents was a 
persistent theme in Berners-Lee's vision for a universal information space. On the most 
fundamental level, users would be able to add new links from one node to another as they 
browsed. 23 This simple notion would allow users to keep the information stored in the 
link structure of the web up to date as they browsed, and it would be easy to link new 
content into an existing web. In the Information Management proposal, Berners-Lee also 
proposed the need for the management of private links from each node, so that users 
could mark up a node with their own comments in the form of links that would not be 
seen by others. 24 
Berners-Lee elucidated even further on his vision of collaboration in 1996 in an 


18 Information Management: A Proposal, pg. 10 
19 Weaving the Web, pg. 19 
20 Weaving the Web, pg. 16 
21 Weaving the Web, pg. 16 
22 Weaving the Web, pg. 16 
23 Weaving the Web, pg. 16 22 
22 Page 23 24 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
23 
interview published in the MIT Technology Review. He envisioned web documents as 
fully collaborative living documents, able to be edited in real-time by several people 
simultaneously and include live video and/ or audio feeds from each participant: 
BERNERS-LEE: Say that you conduct a meeting as a hypertext document. You start by dragging in a video version of yourself, with real-time sound. 


You remind those invited to come by sending them a hypertext e-mail with a pointer to the meeting. To join, they just follow the link. They can not 
only read this meeting/ document, but they also write to it. Some join by audio and some drag their own video into the document. People introduce 
points by writing them into the minutes, making links to background material. At one point in the meeting three people realize they need to 
discuss something separately, and with a single keystroke one forks off a new meeting document that they will catch up with later. 25 


Clearly, Berners-Lee envisioned a large degree of two-way interaction on the Web, 
especially for the purpose of facilitating group communication. 
Above all, Berners-Lee desired that a simple, straightforward design with lots of 
room for future expansion. Too much innovation and new concepts all at once would 
result in lower acceptance for his vision. Berners-Lee's decision to keep his vision simple 
and pragmatic allowed others to quickly adopt the Web into their own software efforts. 


From Vision to Proposal 
Berners-Lee wished to implement his vision as a new information management 
system for CERN. Mike Sendall, Berners-Lee's boss, asked him to write a proposal for 
the system, which was entitled "Information Management: A Proposal". This proposal 
encompassed most of Berners-Lee's vision for a universal information space, and it 
especially accentuated the aspects that were most important for the scientists and 
management at CERN. A large section of the proposal is dedicated to explaining how 
traditional hierarchical structures like those used in older information systems at CERN 
were inflexible and therefore forced unnecessary constraints onto information stored 
within these systems. Berners-Lee observed that organizations and relationships at 
CERN often did not follow rigid hierarchical structures, especially when looked at over 
time. Information about a project should be able to grow and evolve along with the 


24 Information Management: A Proposal, pg. 10 23 
23 Page 24 25 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
24 
organizations that work on it, and therefore "the method of storage must not place its own 
restraints on the information". 26 
The proposal also touched upon several other issues important to CERN, 
including specific examples of existing network-based services that would be converted 
over to Web services, such as the phone book and various help systems. Berners-Lee 
also mentioned the possibility of making the system work over DECnet, as CERN's 
computer population had a significant percentage of DEC VAX mainframes, which 
primarily used DECnet. And finally, the proposal briefly included an overview of the 
project management of developing the system, which was estimated at 2 people's time for 
6-12 months, and then further effort to get CERN's existing systems online in the new 
system. 


The Engineering Realities 
Having created a vision of a universal information space that allows both viewing 
and cooperative editing of hypertext documents, Tim Berners-Lee faced the challenge of 
turning his ambitious vision into an engineering reality. Berners-Lee's first thought was 
to add networking capabilities to existing hypertext systems. At the European 
Conference on Hypertext Technology in 1990, Berners-Lee approached Ian Richie of 
OWL Ltd. about merging his ideas with OWL's Guide hypertext system. The Guide was 
"astonishingly like what [Berners-Lee] had envisioned for a Web browser," including the 
functionality to edit hypertext documents. 27 However, Berners-Lee was unable to capture 
Ritchie's interest. Ian Richie, the founder of OWL, had a different view. OWL's Guidex 
already had most of the features, including retrieval of documents over a network, that 
Berners-Lee wanted for the Web. What OWL did not realize was the fact that the 
networking aspect of an information system would prove to be more important than its 
hypertext aspects. Dynatext, created by Electronic Book Technology, was another 
potential hypertext base for Tim Berners-Lee's Web, but here too, he was unable to make 


25 Technology Review June 1996 
26 Information Management: A Proposal, pg. 4 
27 Weaving the Web, pg. 26 24 
24 Page 25 26 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
25 
his case. Berners-Lee attributed the lack of enthusiasm for his ideas in the hypertext 
community partly to the lack of a good demonstration of the Web. He felt that 
"explaining the vision of the Web to people was exceedingly difficult without a Web 
browser in hand". 28 It was becoming clear that if the Web were ever to see the light of 
day, Tim Berners-Lee would have to design and build it himself, from scratch. 
The first implementation of the Web was completed in about two months by Tim 
Berners-Lee, working alone. Berners-Lee had developed a graphical browser/ editor that 
could read and write HTML documents through HTTP. He also developed the first Web 
server at info. cern. ch to serve up the documentation and his notes for HTTP and 
HTML. 29 Early HTML and HTTP were extremely simple out of necessity. Being the 
only developer, Berners-Lee could not afford to implement a complex set of 
functionalities. The HTTP protocol is therefore layered on top of existing Internet 
standards such as TCP/ IP (Original HTTP, 1991). For example, when a URL such as 
http:// web. mit. edu/ index. html is entered into a Web browsers, the HTTP connection takes 
place in three phases. First, the type of the URL request is determined to be HTTP. Then 
the name of the remote server, web. mit. edu, is extracted from the URL and the actual IP 
address is looked up via the DNS system. Since no TCP port number on web. mit. edu has 
been specified, the default port for HTTP connections (port 80) is used. The Web 
browser now makes an ordinary TCP connection to port 80 on the computer named 
web. mit. edu. Once the connection is established, the Web browser sends a single line of 
ASCII characters, which is the actual HTTP request. In our example the line will read 
"GET index. html." The HTTP server running on web. mit. edu receives this request and 
tries to find the file "index. html." If the file exists and is readable by all users, then it is 
sent to the Web browser. But if for some reason the file cannot be located or cannot be 
read for some other reason, an error message is sent to the Web browser instead. When 
the server has finished sending either the file or the error message, it closes the TCP 
connection with the Web browser, thus ending the HTTP connection. It is then up to the 
Web browser to correctly display the document that it has received. 
As we can see from the above example, the original HTTP design was a "non-


28 Weaving the Web, pg. 27 
29 Weaving the Web, pg. 29 25 
25 Page 26 27 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
26 
computer-scientist's protocol" that favored simplicity over efficiency. 30 Its most 
important feature, however, was that it worked well enough to be a usable system. 
Similarly for HTML, simplicity and functionality were its strengths. The original HTML 
specification contained little more than a dozen distinct markup tags. Only the most 
rudimentary document markups were supported: paragraphs, lists, highlighting of text, 
headings, indices, glossaries, and of course hypertext links. In comparison, HTML 4.01 
defines 91 distinct tags and literally hundreds of additional attributes that can be used to 
modify the behavior of these tags. All of the tags in HTML 4.01 are useful today, but if 
Berners-Lee had made the very first version of HTML just like HTML 4.01, chances are 
the Web would not have taken off at all. 
The simplicity of HTML and HTTP also had their drawbacks, however. One of 
the features Tim Berners-Lee had to sacrifice was the idea of cooperative editing of 
HTML pages over the Web. In order to implement such a two-way Web, the HTTP 
protocol would have needed a good authentication system so that users can be given 
permission to write HTML documents onto a remote computer. The 1992 version of 
HTTP shows that some provision was made for such an authentication system. In the 
HTTP request specification, one of the fields is named "Authorization," and its purpose 
was to allow the originator of the request to authenticate himself or herself to the HTTP 
server. Once a user is authenticated, he or she can be given permission to changed 
HTML documents on the server using the "POST" and "PUT" methods. "POST" and 
"PUT" are similar to the "GET" method, and they allow the user to amend and create files 
on the server, respectively. These features, if they were implemented, could have 
supported Berners-Lee's vision of a two-way Web. But the "Authorization" field was 
designated at To Be Specified (TBS) in 1992, and even today, there is still no simple way 
to directly modify existing HTML pages on the Web. It is interesting to note that the 
W3C's Amaya project is still striving for Berners-Lee's ideal of a two-way Web by 
integrating browsing and editing functionalities into a single Web client. However, for 


30 Dave Gifford interview. 
31 Pei Wei email interview. 
32 Pei Wei email interview. 
33 Thomas Bruce interview. 
34 Thomas Bruce interview. 
35 An Interview with Tim Berners-Lee (video). 26 
26 Page 27 28 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
27 
the foreseeable future, most HTML documents on the Web will remain read-only. 
Despite the shortcomings of the first implementation of the Web, Berners-Lee 
remained steadfast on his altruistic goal of building a universal information space for the 
entire world. In an email to Dan Connolly on November 8, 1991, Berners-Lee 
summarized the vision of the Web saying, "I hope I have clarified the W3 team's 
philosophy, and perhaps convinced you to contribute, to our mutual (and the world's) 
benefit." 36 


The Struggle For Acceptance 
When Tim Berners-Lee implemented the first version of the Web, the world did 
not beat a path to his door. Just as Thomas Kuhn predicted that new scientific paradigms 
are met with much resistance, the World Wide Web met its share of resistance. Two 
important sources of early resistance to the Web were CERN and the hypertext 
community. It is a common misconception, perpetuated to a degree by Tim Berners-Lee's 
own accounts, that neither CERN management nor the hypertext community saw 
the usefulness of the World Wide Web. It is true that they underestimated its </P>
    <p> *******************************************************************************************</p></div>
<p align="center">&nbsp;</p>

ance, 
as almost everyone did at the time. But CERN and the hypertext community were not 
blind to the potentials of a networked information system. Then rejected Tim Berners-Lee's 
project, which is quite different from rejecting the premises of the World Wide 
Web. We will explore some of 
the reasons for CERN and the 
hypertext community to reject 
the early Web. 


CERN 
As we have said earlier, 
CERN is the center of 
European high-energy physics 


36 Email from Berners-Lee to Dan Connolly, included as Item 1 in Appendix A. 
Figure 6 Particle accelerator at CERN. 27 
27 Page 28 29 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
28 
research. As such, it had a tremendously diverse culture and a great deal of complex 
social needs. For example, the people working at CERN were not all working at CERN. 
Many of the scientists involved in CERN projects were scattered all over Europe, the 
United States and even Japan. There were also a large number of projects going on at the 
same time. Such a heterogeneous environment cried out for a networked information 
system. Tim Berners-Lee, as a software consultant for CERN, was keenly aware of this 
need. In his numerous proposals to CERN management, Tim Berners-Lee always 
stressed the usefulness of his World Wide Web project to the people at CERN. CERN 
was of course also well aware of its own needs, and it seems puzzling that they did not at 
least explore Tim Berners-Lee's ideas more fully. 
However, the puzzle is much less puzzling when we consider the fact that 
CERN's primary focus is physics research. CERN's policy regarding software systems is 
"buy, not build." 37 However, this policy alone is not the only reason CERN did not take 
up Tim Berners-Lee's World Wide Web project. According to David Williams, Tim 
Berners-Lee focused a lot of his persuasive efforts on the hypertext aspects of the Web. 
However, there were a number of people at CERN, Williams included, who did not feel 
that hypertext was an appropriate way to represent information. Williams also thought, 
"Explaining his ideas is not [Tim Berners-Lee's] real strength." 38 As we can see from 
FIGURE 7, Tim Berners-Lee's diagram of how the Web works, 40 presents quite a 
complex picture of his ideas. This complex picture presents a direct contrast to the 
relatively simple implementation of the Web. This disparity between the simplicity of 
the Web's implementation and the complexity of Berners-Lee's explanations may explain 
why CERN hesitated to back a full-scale project to develop the Web. As a physics 
research institution, CERN had to devote most of its resources to that end. However, if 
Tim Berners-Lee had presented the Web as a less daunting project, it is possible that 
CERN might have undertaken to support his efforts more. 


37 Weaving the Web 
38 David Williams email interview. 
39 Information Management: A Proposal 
40 Information Management: A Proposal 28 
28 Page 29 30 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
29 
Figure 7 Tim Berners-Lee's flow chart for the Web 
The Hypertext Community 
In trying to minimize the development efforts required for the World Wide Web, 
Tim Berners-Lee had approached several hypertext system builders about possible 
collaboration. Berners-Lee believed that once people saw his simple implementation of 
the Web, they would realize the enormous potential of a global information space. 
However, his early implementation of the Web was not an impressive system to people 
within the hypertext community. At the Hypertext '91 conference in San Antonio, Tim 
Berners-Lee demonstrated the WorldWideWeb browser on his NeXT machine to the 
attendees. However, as a hypertext system, HTML was far too simple to make for a 
convincing demo. 41 The key to Berners-Lee's idea was of course the marriage of 
hypertext and the Internet. But to the hypertext community, his implementation was not 
an interesting hypertext system because it addressed none of the issues facing hypertext 
systems of the time. 
For example, link consistency was a big concern for most hypertext systems in the 
early '90s. When contents move in a system, links that point to the contents can become 
broken. Many hypertext systems were designed to deal with this problem. The Web on 29 
29 Page 30 31 
THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
30 
the other hand, completely ignored the problem of link consistency. Tim Berners-Lee 
was aware of the link consistency problem, but he correctly recognized that there was no 
simple, scalable way to ensure global link consistency in a worldwide information 
system. However, by ignoring the interesting hypertext research problems, Berners-Lee 
was not able to persuade the hypertext community to help him develop the Web. 


Why the Web Won 
Despite the resistance Tim Berners-Lee faced at CERN and in the hypertext 
community, the Web did beat out its competitors to become the most widely used 
electronic information system. The triumph of the Web comes from a combination of 
many factors. Tim Berners-Lee's promising vision, the simplicity and openness of the 
Web's design and initial implementation, the tremendous grass-roots support from users 
and developers, and fortuitous timing all contributed to the eventual success of the Web 
as a new paradigm. 
Tim Berners-Lee, though he may have erred in presenting too grand a vision to 
CERN, received an entirely different reaction from the world at large. For people outside 
CERN, the idea of a global information space was very appealing. However, the 
simplification of Tim Berners-Lee's vision was the only way for the Web to take roots 
and grow. The simplicity of early HTML and HTTP not only made them easy to 
implement but also easy to extend incrementally. Unlike more mature hypertext systems 
such as Dynatext, which had consistency checks to ensure their hypertext documents 
were valid, HTML was extremely lax in terms of error checking. While it created 
something of a problem because ill-formed documents could not be viewed, the lack of 
error checking was the most important factor in the growth of the Web. Since there is no 
real way to ensure that an HTML document conformed to the specifications given by 
Berners-Lee, browser implementers were free to add their own tags to HTML and have 
them be appropriately interpreted by their own browsers. All a browser implementer had 
to do was let his users know that a new tag was supported by his browser. For example 


41 Peter Brown phone interview. 30 
30 Page 31 32 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
31 
the developers at NCSA added tags to display graphics to their browser implementation 
in just such a unilateral fashion. 
When Tim Berners-Lee released his creation to the people, the Web enjoyed 
tremendous grass-roots support on the Internet. Berners-Lee had always known the 
</P>
    <p> *******************************************************************************************</p></div>
<p align="center">&nbsp;</p>

ance of gaining a "critical mass" of users and information on the Web. In fact, the 
first Web server that he set up at info. cern. ch served documentation on the Web itself and 
instructed people on how they can set up their own Web servers. 
Here too, the simplicity of HTTP and HTML were of great value. Early HTTP 
was so simple that an HTTP server could be implemented as a UNIX shell script. 
Similarly, the HTML defined by Berners-Lee was simple enough that several individuals 
were able to develop Web browsers from scratch. Pei Wei, then at the University of 
California at Berkeley, developed Viola, a graphical browser for the X Windows 
environment (Figure 8). Wei 
commented on the "simple and elegant" 
URL scheme used by HTTP that 
attracted him to the Web. 43 Wei 
believes that if HTTP and HTML were 
not "plainly easy for a developer to 
work with," then the Web would have 
lost that developer and "that much of a 
resource for this grass-root supported 
system". 44 Other developers that 
created early browsers such as Tom 
Bruce, whose Cello browser was the 
earliest Windows based browser, also 
credit the simplicity of the Web for its 
quick acceptance. Bruce believes that Berners-Lee was able to "avoid letting the perfect 
become the enemy of the good" by keeping early HTML and HTTP simple and adding 


42 Mark McCahill phone interview. 
43 Pei Wei email interview. 
44 Pei Wei email interview. 


Figure 8 ViolaWWW Screenshot. 31 
31 Page 32 33 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
32 
more functionality as the Web grew. 45 However, the Windows platform was largely 
ignored by the early Web enthusiasts. Even Berners-Lee's cross platform library, 
libwww, worked poorly under Windows because of memory allocation issues. Thomas 
Bruce, in the course of writing Cello, rewrote the entire libwww for Windows. 46 For a 
lone developer, this was a great deal of work. But because the Web was a simple enough 
and exciting enough, it spread to all the important platforms through the work of its 
grass-roots supporters. 
Once free Web browsers became available on all platforms and independent 
HTTP servers began to come online, the Web took off quickly. Tim Berners-Lee 
observed that since 1991, the Web has been growing at an exponential pace. 47 The 
advent of cheap PCs and a commercialization of the Internet further fueled its incredible 
growth. In less than 10 years, the Web went from a system no one wanted to a household 
word thanks in large part to the grass-roots support it enjoyed in its early days. 
The World Wide Web had some additional advantages over its closest competitor, 
Gopher. In the Spring 1992 issue of the Electronic Networking journal, Berners-Lee 
wrote that since Gopher uses the "directory and file model to implement a global 
information system," it would "map into the Web very naturally, as each directory 
(menu) is represented by a list of text elements linked to other directories or files 
(documents)." 49 In other words, the Web was a more general system that is capable of 
encompassing most of Gopher's capabilities. Another advantage the Web had over 
Gopher was that the "web gave information away and had ads to support it." 52 Companies 
saw the opportunity to make money and therefore chose the web over Gopher. The Web 
was "very pretty" and while librarians loved Gopher for its orderliness, companies loved 
the Web for its graphics capabilities. 
While the graphical capabilities of Mosaic did much to popularize the Web, such 


45 Thomas Bruce interview. 
46 Thomas Bruce interview. 
47 An Interview with Tim Berners-Lee (video). 


49 Electronic Networking, pg. 56 
51 Mark McCahill phone interview. 
52 Mark McCahill phone interview 
53 Thomas Bruce phone interview. 32 
32 Page 33 34 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
33 
proprietary extensions were a double-edged sword. On the one hand, they cause 
incompatibilities among different browser implementations, thus dividing Berners-Lee's 
vision of a universal information space. An HTML document containing Mosaic specific 
tags would not display correctly on any browser except NCSA Mosaic. On the other 
hand, proprietary extensions were quick and direct ways for enterprising developers to 
add new capabilities to the Web. In the early days of the Web, the benefits of these 
extensions outweighed the drawbacks. The graphical capabilities of Mosaic were one of 
the main reasons the Web became so attractive to people outside the circles of academics 
and research. As the saying goes, "graphics sells," and in the case of the Web, the 
graphics gave it a tremendous leg up on competing systems such as Gopher. 54 However, 
as the Web grew and Mosaic became the dominant browser, the danger that a group of 
aggressive developers could hijack the development of the Web became increasingly 
present. It was clear that in order for Berners-Lee's vision of the universal Web to 
survive, some neutral body was needed to build consensus among the various forces that 
were pulling the Web in different directions. 


Adoption of the New Paradigm 
The Web needed to leave CERN for it to truly succeed -it was, after all, a particle 
research facility and not focused on computer science research. Berners-Lee therefore 
moved his work and advocacy for the Web to the Massachusetts Institute of Technology's 
Laboratory for Computer Science and started the World Wide Wed Consortium (W3C). 
"The W3C was founded in October 1994 to lead the World Wide Web to its full potential 
by developing common protocols that promote its evolution and ensure its 
interoperability." 55 Michael Dertouzos, the Director of the Laboratory for Computer 
Science describes that there was "tremendous synergy" between him and Berners-Lee 
when they met in Zurich in 1994. Dertouzos was interested in creating an "information 
marketplace" and saw the world-wide web as the "potential underlying mechanism" for 


54 Mark McCahill phone interview. 
55 http:// www. w3c. org web page. 33 
33 Page 34 35 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
34 
his vision. 56 MIT had just successfully spun off the X Consortium and this provided 
good timing for Berners-Lee to set up the W3C when he was invited to do so by 
Dertouzos. He saw it as his opportunity to continue to drive the future of the Web as it 
grew. CERN allowed the Web to leave and gave up its rights to the technology that 
Berners-Lee had developed there. David Williams, a manager at CERN, credits this 
happening to Berners-Lee and Robert Cailliau. He says, "Tim was strongly of the opinion 
that only an open release would allow it [the Web] to take-off. While we are all happy 
with what happened I tend to feel that CERN should have tried to get a little more public 
recognition for our/ his work." 57 But, CERN did not try to do that, and the Web took off 
because of it. 
The W3C is very active today in promoting standards for different web 
technologies. Companies pay a fee for membership and, while they are not forced to 
abide by the recommend standards that are developed by the W3C, they are incompatible 
with others if they chose not to. The aforementioned "browser wars" in which Mosaic 
and other browsers battled to outdo each other by adding more and more proprietary 
features was brought under control by the consensus-building approach of the W3C. This 
stabilization and standardization fostered the growth of the Web, as developers and 
companies can depend on there not being major differences and incompatibilities 
between the different browsers. 
The birth of the W3C signifies the acceptance of the Web as a new paradigm in 
information sharing. The W3C is a forum for the building of consensus on the direction 
of the Web. It also represents the community that lives and works within this engineering 
paradigm. By working to extend the Web and fulfill its potential, the work of the W3C 
represents the kind of normal engineering that goes on within an established paradigm. 
Once people began to accept the W3C as an authority in matters regarding the Web, the 
engineering revolution has truly come full circle. 


56 Michael Dertouzos interview. 
57 David Williams email interview. 34 
34 Page 35 36 
THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
35 
Conclusion 
We have tried to tell the story of the Web in terms of an engineering revolution. 
The progression from the existing paradigm consisting of unconnected computers and 
isolated islands of information to a connected world of universal information sharing is 
truly a revolutionary development. We have shown the similarities and differences 
between Kuhn's model of scientific revolutions and engineering revolutions. 
People today do not give a second thought about looking for information on the 
Web. While it may not always be obvious as to how to find the information, it's almost 
guaranteed that it exists on some Web page. People just 10 years ago looked for their 
information in a dramatically different way: the world has truly experienced a paradigm 
shift. 35 
35 Page 36 37 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
36 
Bibliography 
Books 
Berners-Lee, Tim, Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web by Its Inventor. HarperCollins 1999 


Gilles, James, and Cailliau, Robert, How the Web was Born. Oxford University Press, 2000 
Naughton, John. A Brief History of the Future: The Origins of the Internet. Weidenfeld: London, 1999. 


Papers and Journal Articles 
Berners-Lee, T. J., Cailliau, R. and Groff, J.-F. "The World Wide Web." Computer and Networks and ISDN Systems 25. 1992: pp454-59. 


B. Kahle et al., "Wide Area Information Servers : An Executive Information System for Unstructured Files", in: Electronic Networking: Research, Applications and Policy, Vol. 
2, No. 1 (Meckler, New York, 1992) 59-68. 
Berners-Lee, T. J. et al., "World-wide web: the information universe", in: Electronic Networking: Research, Applications and Policy, Vol. 2, No. 1 (Meckler, New York, 
1992) 52-58. 
Berners-Lee, T. J. et al., "World-wide web: an information infrastructure for high-energy physics", in: D. Perret-Gallix, ed., Proc. International Workshop on Software 
Engineering and Artificial Intelligence for High Energy Physics, La Londe, France, January 1992. 


McCahill, M. et al., "The Internet Gopher: An Information Sheet", in: Electronic Networking: Research, Applications and Policy, Vol. 2, No. 1 (Meckler, New York, 
1992) 67-71. 
Berners-Lee, Tim. "Information Management: A Proposal." CERN, 1989. 
Berners-Lee, Tim and Cailliau, Robert. "World Wide Web: Proposal for a HyperText Project." CERN, 1990. 


"The Web Maestro: An Interview with Tim Berners-Lee", Technology Review, July 1996, archived at http:// www. techreview. com/ articles/ july96/ bernerslee. html 36 
36 Page 37 38 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
37 
Roush, Wade, "MIT Reporter: Spinning a Better Web", Technology Review, April 1995, archived at http:// www. techreview. com/ articles/ apr95/ BernersInterview. html 
Interviews 
Email interview with David Williams, Computing and Networks Division CERN, November 11, 2000 


Interview with Alan Kotok, Associate Chairman of the World Wide Web Consortium, November 13, 2000 
Telephone interview with Mark McCahill, Computer and Information Services, University of Minnesota, November 15, 2000 
Email interview with Thomas Bruce, Legal Information Institute at Cornell Law School, November 15, 2000 
Interview with Dave Gifford, Programming Systems Research Group, MIT Lab for Computer Science, November 15, 2000 
Telephone interview with Peter Brown, University of Kent, November 17, 2000 
Telephone interview with Ian Ritchie, Office Workstations Limited (OWL), November 20, 2000 


Email interview with Pei Wei, University of California Berkeley/ O'Reilly & Associates, November 21, 2000 
Telephone interview with Doug Engelbart, Stanford Research Institute/ Bootstrap Institute, November 27, 2000 
Interview with Michael Dertouzos, Directory of MIT Lab for Computer Science, November, 28, 2000 


Videos 
A conversation with the man who invented the World Wide Web [videorecording] : Tim Berner-Lee [sic], 1996, MIT Archives collection 


Demonstration of NLS, Doug Engelbart, http:// sloan. stanford. edu/ MouseSite/ 1968Demo. html 


Web Sites 37 
37 Page 38 39 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
38 
RFC1630: Universal Resource Identifiers in WWW Http:// www. ietf. org/ rfc/ rfc1630. txt? number= 1630 
Archives of the www-talk forum Http:// ksi. cpsc. ucalgary. ca/ archives/ WWW-TALK/ archives. html 
List of Original HTML Tags, 1991. http:// www. w3. org 
Berners-Lee, Tim. HyperText Transfer Protocol Design Issues, 1991. http:// www. w3. org 
Grahn, J. Harmon. "The World Wide Web." http:// www. olypen. com/ harmon/ np304. htm 38 
38 Page 39 40 


THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
39 
Appendix A: Selected Emails from the WWW-Talk archive 
All of these emails are hosted at http:// ksi. cpsc. ucalgary. ca/ archives/ WWW-TALK/ archives. html. 


Item 1 – Email from Tim Berners-Lee to Dan Connolly 
Re: Motif browser status timbl (Tim Berners-Lee) 


Date: Fri, 8 Nov 91 13: 35: 26 GMT+ 0100 
From: timbl (Tim Berners-Lee) 
To: connolly@ pixel. convex. com 
Subject: Re: Motif browser status 
Cc: kharris@ pixel. convex. com, www-talk 


Dan, 
Thanks for your message. Obviously you know what you are doing with 
X11 browsers -we are impressed by what you have done to date. I was 
interested to hear that you are working on AVS -I have had some 
contact with AVS people at UNC. 


You make a good point that the world has been waiting for a good 
formatted text widget under Motif. One exists under NeXTStep, Robert 
Cailliau is just adapting one for the Mac for hypertext, but under 
Motif it has been lacking. Of course, hundreds of people have 
written them: all the word processors have them in, and products like 
dynaText, etc. However, there is none in the public domain. 


CERN like Convex has a copyright on all code, but we are doing our 
best to release W3 code as widely as possible, and possibly overcome 
this limitation. Why? 


The concept of the web is of universal readership. If you publish a 
document on the web, it is important that anyone who has access to it 
can read it and link to it. In order to make this possible, we don't 
need very new technology --what we do need is 


1. A common open naming/ addressing format 
2. Sufficiently powerful underlying protocols 
3. Sufficiently powerful data formats 
4. Some free implementations 


Now we have defined the (1), which did not exist before. We have 
supplemented the (2), where some protocols do exist. We have added a 
little to (3) though we will use all existing and new formats. We 
have written some code. 


You say your work would be of considerable valuer to convex. Yes, 
that is true. You must ask yourself whether it would be of more value 
to convex if kept private or released for general consumption. If you 
release it, 39 
39 Page 40 41 
THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
40 
-Convex gets the credit and a higher profile, 
(as Thinking Machines has with WAIS indexers for example). 


-Anyone in the world can read the information you supply 
with the same tool as they use for other information. 


-You get a lot of useful feedback from users on the network 
-A lot of people would be able to profit from what you have 
done 


You have to compare this scenario with that if you keep the code 
private. You will be able to use it internally. Would convex be able 
to profit from by selling it? If so, how many people would actually 
buy it? Will the AVS project benefit from a closed private 
documentation scheme? 


On these grounds alone, you may conclude that it is in Convex's 
interest to release the code. Still, you ask what we can "put on the 
table". If it would make it easier to justify the release of code, 
we would be happy to make all CERN-developed W3 code officially 
available to Convex under a more or less formal joint project 
agreement. Note that we are producing a parallel set of parsers and 
access mechanisms for HTML, newgroups, WAIS, prospero, etc. We have 
gateways, and other browsers. The line-mode browser you know, the Mac 
one is coming along, we may have a full-screen character grid browser 
too. We are currently unifying the browser architecture so that all 
access mechanisms can be used by all browsers. I'm not sure that 
either of our sides would want to be contractually bound to produce 
or maintain anything -the agreement would be just as-is code sharing 
of what exists when it exists, no strings. 


You ask about graphics. That cannot be our next priority, as we need 
to get the new architecure and general format negociation worked out. 
In many cases, we find that there are GIF/ TIFF viewers on various 
platforms, and one can link in to them. We don't want to make a new 
graphics file format a la Mac/ PICT, but we are intrerested in 
conversion code. Have you heard of editable Postscript? That might be 
what you are looking for. (See 
http:// info. cern. ch/ hypertext/ Standards/ PostScript/ IPF. html) 


I don't know whether your company has a mechanism for allowing code 
to be released into the public domain (or General Public License). If 
it is politically impossible, then that's a pity. (We do have a 
group of students in Finland working on an X implementation, and if 
that doesn't work out we could write it ourselves. It may also be 
that more that one implementation with a different style will be 
interesting. Obviously it would be rather a duplication of effort, 
though we are under a lot of pressure from our management and users 
to put this at the top of the agenda.) 


I hope I have clarified the W3 team's philosophy, and perhaps 
convinced you to contribute, to our mutual (and the world's) benefit. 


Tim 
PS: Yes, I think you ought to be on www-talk, Dan. I'll put you on. 40 
40 Page 41 
THE STRUCTURE OF ENGINEERING REVOLUTIONS – THE WEB AS AN ENGINEERING PARADIGM 
41 
The traffic is not too high. 
__________________________________________________________ 
Tim Berners-Lee timbl@ info. cern. ch 
World Wide Web project (NeXTMail is ok) 
CERN Tel: +41( 22) 767 3755 
1211 Geneva 23, Switzerland Fax: +41( 22) 767 7155 41 

Page Navigation Panel 
1 2 3 4 5 6 7 8 9 
10 11 12 13 14 15 16 17 18 19 
20 21 22 23 24 25 26 27 28 29 
30 31 32 33 34 35 36 37 38 39 
40 41