After Tim Berners-Lee successfully graduated from Oxford, he moved to the CERN research centre in Switzerland, where he was offered a position as a software engineer. Before he decided at CERN to solve the problems of information sharing once and for all, he worked on several minor projects. Berners-Lee was, for example, called upon to create more user-friendly controls for a vacuum control system[1]. Instead of displaying the entire control interface of the system at once on a screen of 24 lines — with a limit of 64 characters on each of them — he decided to create something we would call a “menu”: on the main screen he left only a display of the main modules of the system, and only after clicking on a selected part did the operator get to a “subpage” where the relevant components were available.

While the project mentioned above contained no elements of hypertext (except perhaps the possibility of displaying another part of the document by clicking), the first project implementing the hypertext model was ENQUIRE from 1980. ENQUIRE made it possible “to structure information in any arbitrary way. It did not have to be displayed in a tree structure or in a set of tables”[2]. ENQUIRE offered bidirectional links, a functionality that is missing, for instance, from today’s form of the Web platform: a page that is linked to does not know that it is being linked to. According to Tim Berners-Lee[3], the main limiting factor of this project was the impossibility of linking to external resources; ENQUIRE supported links between files as well as between parts of a single file. But all of that within a single file system. If Berners-Lee wanted to create a truly global information system, he had to get one more chance; that chance was the World Wide Web.

Which of the projects listed so far influenced Berners-Lee in the design and implementation of the components of the Web? He was indirectly influenced by Ted Nelson and his Literary Machines; he considers Douglas Engelbart’s NLS design to be the closest to the Web[4]. We learn that around 1989 hypertext was already a generally known technology; SGML, in turn, was used extensively at CERN itself to structure internal documents[5]. It seems that Berners-Lee, thanks also to the testing of and feedback on the ENQUIRE project, had much of what he needed to start working on what is today the most widely used service of the Internet — the World Wide Web service.

The “World Wide Web” project, created in 1990, comprised the HTML markup language, the HTTP hypertext protocol, the URI mechanism for addressing documents, and a browser. The first web browser, designed by Berners-Lee, was called “WorldWideWeb” and offered a unique feature: unlike today’s browsers, it was conceived as a browser and an editor at the same time; using this graphical WYSIWYG[6] tool, users could thus edit pages directly while browsing them[7]. We carry out a more detailed analysis of the Web platform in the section The Web as a Technological Platform.

The development of the first web browsers

The first web browser — WorldWideWeb — was designed for the NeXT system, whose performance exceeded the computing power of the machines found in ordinary households. Then, in 1991, came the first multiplatform browser (the second overall), the Line Mode Browser, written by Nicola Pellow. It offered only a text interface, and therefore all control (moving to another page) had to be done by entering the appropriate command. In the same year, the Erwise browser, developed by a Finnish university team, was released. Erwise is usually described as the first browser with a graphical interface[8]; however, as we have seen, this is not the case, for the first graphical browser was that very first browser developed by Berners-Lee. This mystification most likely stems from the fact that Erwise does hold a certain primacy: it was the first browser for the X Window System (X11). A competing browser for X11 was Viola WWW. Pei-Yuan Wei, the author of Viola WWW, drew inspiration from Apple’s HyperCard project. Like Erwise, it also offered support for multiple fonts, highlighting of links, and back and forward buttons, including a button for going to the home page. The user could “clone” individual pages into new windows. Viola also offered support for recording visited pages and the option of bookmarking a page.[9] Viola integrated groundbreaking support for scripting, embedding applets in the body of a page, its own version of style sheets and support for the HTML 3.0 standard.[10]

Viola is an example of how software with a longer list of features on paper is no guarantee of success in competition with less advanced products. For it is above all users who decide on success; from the history of design movements we have learned that usability, user-friendliness, attractiveness and even a steeper learning curve are often more important factors for users. The legendary Mosaic browser, created by Marc Andreessen and Eric Bina at NCSA[11], is considered not only the browser that made the Web accessible to the masses, but also the browser that brought the Web platform closer to graphic design and DTP[12]. Mosaic was, after all, the first browser that made it possible to display images directly in the document of a page. All the previous browsers dealt with images by displaying them in separate windows; and mostly only after the user had clicked on a link to display them. With images integrated directly into the document, including text wrapping around them, the Web, in the words of Tim Berners-Lee[13], suddenly became a more attractive and seductive place. If we were to place Mosaic in the context of the field of web design, one could argue that Mosaic stands at the genesis of web design. The combination of images, early style sheets, the possibility of using different typefaces and, last but not least, text formatting is, after all, already the domain of the traditional design disciplines mentioned, which we know from the world of print. Mosaic gave the creators of web pages room to raise pure utility to an aesthetic experience evoked by a fitting arrangement of the available media elements. Add to this that Mosaic was ported from the original Unix code to the popular Mac and PC platforms. Berners-Lee writes that “[…] Mosaic was the simplest step for a beginner to get onto the Web, and therefore a critical element in the explosion of the Web”[14].

After disagreements at NCSA, Marc Andreessen and his colleagues founded the Netscape Communications Corporation and in 1994 came onto the market with a new (likewise legendary) browser, Netscape Navigator. It did not, however, hold an outright monopoly on the market, since several companies, Microsoft among them, licensed the source code of Mosaic, on the basis of which Microsoft released its own browser, Internet Explorer[15], which in version 1.0 was shipped in the separately sold Microsoft Plus! package for Windows 95. Netscape and Internet Explorer became the dominant browsers on the market and competed for the attention of users. While Microsoft could ship its browser for free as part of its operating system, which was growing in popularity, Netscape had no such option, and so it relied on innovation in developing the technological capabilities of its browser. It was Netscape that introduced to the Web platform such features as frames, cookies and above all the JavaScript scripting language. Microsoft’s answer to JavaScript came in 1996 with its own implementation called JScript.[16]

The confrontation between these browsers intensified, and this period is also aptly called the “browser wars”. Staying competitive and keeping influence on the market spurred developers to further innovation. New technologies and features emerged: improved capabilities of Cascading Style Sheets, the concept of dynamic HTML, or the XMLHttpRequest technology, which Microsoft created for Internet Explorer 5 and which was popularised under the umbrella term AJAX only in 2005, in an influential article by Jesse James Garrett of the consultancy Adaptive Path[17], and in practice, for example, in Google’s applications. Further browser wars went on even after the influence of Netscape declined and its code became the open-source basis of the Mozilla browser: new browsers appeared, Opera, Safari and later Chrome; Internet Explorer, with its dominance of the market and its unwillingness to implement behaviour according to the standards, became the terror and nightmare of web designers, but also of developers who, under pressure from designers and superiors alike, were forced to implement the design of a web page for browsers of disparate and idiosyncratic features and behaviours.

Just as in the case of GML, where various mutually differing implementations required unification and standardisation, which resulted in SGML, browser makers and web developers also began to understand that a certain degree of centralisation and agreement was necessary if “the Web was to reach its full potential”[18]. For these purposes, Tim Berners-Lee founded the W3C consortium as early as 1994. To this day it supports dialogue between individual browser makers and proposes its recommendations for implementing new technologies. History has shown, however, that the work of the W3C was not always in line with the expectations and requirements of browser makers and of the creators of web pages themselves.

Standardisation, the fight over XHTML, W3C, WHATWG

The very first standard was HTML as Tim Berners-Lee designed it. In 1993, that is, a year before the W3C came into being, Berners-Lee submitted a document describing all HTML tags and their functions to the Internet Engineering Task Force (IETF)[19]. After long and fruitful debates on the www-talk electronic discussion list (mailing list), Dave Raggett proposed an improvement of the original HTML called HTML+[20]. This version, however, was never implemented and was instead replaced straight away by HTML version 2.0. It was followed by HTML 3.2, which standardised the use of tables, applets and text flowing around images; it also provided backward compatibility with HTML 2.0.[21] The W3C specification of HTML 4 was finished in 1997. Mark Pilgrim comments on the process by which standards, or rather recommendations, were created as follows:

“HTML has always been a conversation between browser makers, authors [of web pages], standards wonks, and other people who just showed up and liked to talk about angle brackets. Most of the successful versions of HTML have been ‘retro-specs,’ catching up to the world while simultaneously trying to nudge it in the right direction.” [22]

Pilgrim continues:

“Anyone who tells you that HTML should be kept ‘pure’ (presumably by ignoring browser makers, or ignoring authors [of web pages], or both) is simply misinformed. HTML has never been pure, and all attempts to purify it have been spectacular failures, matched only by the attempts to replace it.” [23]

Let us add that all the versions of HTML mentioned so far always showed an effort to preserve backward compatibility. When a new version was introduced, the creators of web pages could use the new features without fearing that existing pages would stop working altogether. The W3C turned away from these characteristic conventions of how recommendations were made in 1997, when it published the recommendation for HTML 4.0 and at the same time announced that it was not preparing another version of HTML; instead, it began to devote itself to XML-based standards.

With this step, the W3C transformed itself from a descriptive organisation into a body attempting to prescribe what HTML, and web pages along with it, should look like. The aim of this shift was precisely the intervention Mark Pilgrim warns against: the purification of HTML. It is true that HTML is a subset of SGML, through which it also inherits several drawbacks: HTML has a fixed set of tags, it cannot be extended with custom modules, and it does not provide the properties that would allow easy machine processing. The W3C decided to intervene and remake HTML into a subset of XML. In January 2000, the W3C published XHTML 1.0. It is clear from the specification that this version brings nothing new; it formulates how to write HTML 4.0 in “XML syntax”, which requires closing tags, lowercase letters and writing the attributes of HTML tags in quotation marks. This version of XHTML specifies the so-called Appendix C, which contains a procedure for writing web pages in the new XHTML syntax while preserving compatibility. Another W3C project from 1999 — Extended Forms (later also XForms) — already gives up on compatibility, for “the goals for the next generation of forms are incompatible with preserving backward compatibility with browsers designed for earlier versions of HTML”[24]. Finally, in May 2001, XHTML 1.1 was released, which, besides minor novelties, required that every XHTML 1.1 web page be sent by the server with the specific MIME[25] header application/xhtml+xml. In other words, if the creator of web pages wanted to build genuine XHTML 1.1 projects, they had to follow that directive and send the content with the given header. The question is: did the creators of web pages actually make use of it?

After the release of XHTML, many creators of web pages resolved that the only “correct” way to create for the Web platform was to use XHTML syntax. This method was understood as dogma. The fact is that the majority of web pages written in XHTML were never XHTML documents. We have mentioned that, according to the specification, every XHTML document had to contain the appropriate MIME header in order to be interpreted as XML. The problem was that the dominant browser of the time — Internet Explorer — could not properly display documents with the application/xhtml+xml header. It should be added that the browsers that were ready to handle XHTML did not display the document either, namely whenever there was a single error in its markup — this, however, was the behaviour required by the specification and was considered an “advantage” of XHTML.

If the Web platform was to fulfil the destiny assigned to it by early pioneers such as Douglas Engelbart — that is, a platform enabling communication and collaboration between people, the sharing of information and the creation of virtual communities — it was necessary to bring users into the life of web pages, not as passive consumers but as active agents participating in the content. According to the XHTML specification, even this content added by users would have to be valid XHTML; every posted comment, every embedded medium would have to pass a thorough check before being displayed, since any invalidity would cause all users to be shown an error message from the parser of the XHTML code. Although it is clear that most pages contained mechanisms for checking the input data from users, few could afford not to display a whole page because of an unclosed tag.

In the end, we thus found ourselves in a situation where many web pages proudly displayed enticing badges announcing the XHTML 1.0 (1.1) validity of the page, yet without the application/xhtml+xml header their content was interpreted by browsers as ordinary text/html content, in other words as HTML 4.0 (4.01). Moreover, the announced XHTML 2.0 was supposed to break away from HTML completely.

In 2004, at one of the W3C conferences, several members led by the Mozilla Foundation and Opera Software expressed their dissatisfaction with the leadership: unlike the other members, they supported preserving backward compatibility and, above all, creating new specifications and new features not prescriptively, but on the basis of what is actually relevant for the developers of web browsers and the creators of web pages in the “real world” of the Web platform. The disagreement led to the founding of the Web Hypertext Application Technology Working Group (WHATWG), backed by the makers of web browsers (except Microsoft). The group began working on several important projects: for the first time, it clearly defined how browsers should process HTML and how they should deal with erroneous markup; it was also responsible for the creation of the HTML5 specification[26]. In 2006, Tim Berners-Lee announced that the W3C and WHATWG would cooperate on further projects. In 2009, work on XHTML 2 was ended and cancelled.

Web 2.0

The absence of backward compatibility and the far too harsh penalty for even a single unclosed tag in the code were not properties of XHTML that matched the direction in which the Web was developing around 2004. New or rediscovered technologies were transforming the Web into a multimedia space whose graphical interface provided “rich user experiences”. Besides richer interaction, web projects began to involve users more in their processes — thanks also to Google’s projects (Google Mail, Google Maps) and others such as Flickr, Amazon, eBay and Wikipedia, the Web turned from a static hypertext project into a dynamic and social space.

These changes and tendencies were summed up in 2004 by Tim O’Reilly under the term Web 2.0 (the term itself had been used for the first time by Darcy DiNucci as early as 1999), a term that has not ceased to be controversial ever since. At first glance, the term may imply that an entirely new version of the Web platform had come into being, but that is not the case. O’Reilly was trying to capture precisely the changes the Web had gradually arrived at. In his much-cited article What is Web 2.0[27] O’Reilly writes that by Web 2.0 he understands: the increased participation of users in the creation of web content, the emergence of wiki projects, and the rediscovery of a set of technologies collectively referred to as AJAX[28], which enable asynchronous communication.

In an environment where Microsoft held a monopoly on the browser market and where the words “mobile web” evoked bad memories of experiences with WAP technology, the creators of web pages could regard designing for the Web as an activity very similar to graphic design in the print medium: the “painter’s canvas” inside the browser had fixed dimensions specified in pixels, and the users’ hardware and software displaying web pages did not differ much within one operating system; and if they did, web creators could still rely on the user having a mouse, a keyboard and a 15” or 17” monitor, and accessing web pages from the static context of the comfort of their home.

It is not hard to understand why this state of the Web platform tempted web creators into thinking that the “right design approach” for the Web was to design web pages so that they always display the same way. Web designers were mostly recruited either from the ranks of engineering-minded creators or from the ranks of designers coming from the offline world of print. In both cases, there is the consistency of properties that is expected of these professions. It was taken for granted that a web page must display the same way on all configurations. There appeared (and indeed still appear today) professionals and design firms whose main motto is precisely this “pixel-perfect” precision.

The problems of this approach began to appear at the moment when Microsoft lost its majority position on the browser market. New browsers were quickly catching up with the popularity of Internet Explorer, which forced creators to stop optimising pages for one specific configuration and to start thinking more generally and more abstractly. As early as 2000, John Allsopp[29] wrote that we need to accept the Web for what it really is: a digital, interactive, and thus new medium. He suggested that flexibility and adaptability are not properties that the creators of web pages should fight against; on the contrary, as Allsopp rightly writes, they are its advantage. But established practices are hard to change, and so this manifesto from 2000 was not met with understanding. That changed in 2007, however.

In 2007, Apple introduced a revolutionary mobile device — the iPhone — which for the first time offered users an experience of the Web that came close to that of using a desktop computer. The device was not perfect, but it contained the Safari browser, which implemented most of the important and modern specifications. At the same time, Apple introduced, for the first time, a device whose touch control was accepted by users as a real alternative to the keyboard and the computer mouse. Thanks to the iPhone, the Web became mobile.

Many other companies followed up on the iPhone with their own products, more or less successfully. Entrepreneurs and web creators alike realised that mobile users were becoming an ever more relevant segment. The first reaction of web creators was the so-called “mobile versions of websites”. These were separate versions available under a special subdomain. After the traditional name of this subdomain — e.g. mobile.domena.cz — these versions were also nicknamed “m dot” sites. The look of these sites was shaped by the dimensions of specific devices; in most cases, by the screen dimensions of the iPhone.

With the growing number of devices, screen sizes and technical capabilities, however, even these mobile versions of websites ceased to be a suitable solution. It was necessary to create pages that were truly flexible and that could be displayed well on as many devices as possible. Web creators thus had no choice but to abandon their conviction that a web page must look the same on every screen. On the contrary, it was now required that a web page be built so that it adapts to the technological and physical limits of the device: it made no sense to ask users to operate the interface of a web page on their smart mobile devices in the same way as on their desktop computers.

The solution came in 2010 from Ethan Marcotte[30], when he introduced the term responsive web design into the practice of web design. As in the case of AJAX, Marcotte did not invent new technologies; he merely combined already existing ones in a fitting way. These are CSS media queries, relative units, and flexible dimensions of images and other media. Web designers gained a methodology for displaying the same content on many devices.

In 2011[31] Luke Wroblewski argued that the creators of web pages should stop regarding mobile devices as secondary and as a necessary evil; he claims that designing for mobile devices can help determine not only the look but also the content of projects. In other words, Wroblewski believes that the natural limits of mobile devices force web creators to design the kind of interface and prepare the kind of content that are truly relevant to users. This gave rise to the mobile first approach to creating web pages as a philosophy in which the limits of mobile devices are used so that web creators can decide what is really important and relevant in their projects. On the other hand, the more powerful and technically capable the device a user uses to access web pages, the more additional content can be made available to them.

Trying to apply and transfer established practices and schemes from the traditional medium of print to the Web platform was not only a failure to understand what the Web as a new medium means and offers, but also a decision that limited the range of possibilities from which the web designer chooses. If web creators are not to repeat the same mistake with mobile devices, they need to keep in mind that mobile devices represent not only a new form of interface, but also a different context of momentary use. These, too, are questions that contemporary web design has to deal with.

A Brief History of the Web — a five-part series:

  1. GILLIES and CAILLIAU, How the Web was born: the story of the World Wide Web, p. 158 ↑
  2. Ibid., p. 170 ↑
  3. BERNERS-LEE, Frequently asked questions ↑
  4. Ibid. ↑
  5. GILLIES and CAILLIAU, How the Web was born: the story of the World Wide Web, p. 160 ↑
  6. An abbreviation of the English phrase What You See Is What You Get ↑
  7. In browsers, however, we do find modules that allow “live” editing of the content of a page. Let us mention, e.g., the Firebug module or Chrome Developer Tools. ↑
  8. At the time of writing this work, the English Wikipedia cites several sources that describe Erwise as the first browser with a graphical interface. ↑
  9. BERNERS-LEE, An updated quick look at ViolaWWW ↑
  10. The Viola World Wide Web Application: 9.1 Features Highlight ↑
  11. National Center for Supercomputing Applications, University of Illinois (US) ↑
  12. Desktop Publishing ↑
  13. BERNERS-LEE, Frequently asked questions ↑
  14. Ibid. ↑
  15. Microsoft and Spyglass kiss and make up ↑
  16. JavaScript: How Did We Get Here? ↑
  17. JAMES GARRETT, Ajax: A New Approach to Web Applications ↑
  18. “Leading the Web to its full potential” is the official motto of the W3C consortium ↑
  19. BERNERS-LEE and CONNOLLY, Hypertext Markup Language (HTML): A Representation of Textual Information and MetaInformation for Retrieval and Interchange ↑
  20. RAGGETT, HTML+ (Hypertext markup format) ↑
  21. PILGRIM, HTML5: up and running, p. 6 ↑
  22. Ibid., p. 8 ↑
  23. Ibid., p. 8 ↑
  24. XHTML™ Extended Forms Requirements ↑
  25. Multipurpose Internet Mail Extensions — specifically, the content-type header indicates the type of content ↑
  26. HTML, Living Standard: 12.2 Parsing HTML documents ↑
  27. O’REILLY, What Is Web 2.0: Design Patterns and Business Models for the Next Generation of Software ↑
  28. Asynchronous JavaScript and XML ↑
  29. ALLSOPP, A Dao of Web Design ↑
  30. MARCOTTE, Responsive Web Design ↑
  31. In 2011 Wroblewski published the book Mobile first, but the first mention of mobile first is found in his web article from 2009, see http://www.lukew.com/ff/entry.asp?933 ↑