It’s now almost 15 years since the launch of Mosaic, the first popular web browser. Since then, the web has changed out of all recognition, but its underlying core architecture of HTML-based text rendered in the cross-platform browser is very familiar. Look 15 years ahead, however, and the web is likely to look very different.
To understand why, and how, you have to understand why the web was invented. Tim Berners-Lee’s original 1989 pitch for the web, “Information Management: A Proposal”, wasn’t a manifesto for a universal publishing medium, but a recipe to enable CERN’s widespread scientists to access and update a centralised pool of shared information. His HTML was a radically simple text markup language with about 20 structural elements such as hierarchical headings, ordered and unordered lists, quotes and addresses, aimed at making it child’s play for authors to use.
Its intrinsic simplicity and focus on text content had huge implications for early web design: there wasn’t any! As the proposal’s conclusion put it: “We should work toward a universal, linked information system, in which generality and portability are more important than fancy graphics techniques.” HTML had virtually nothing to say about presentation, devolving all decisions about how tags should be displayed to the browser. Other markup languages such as TeX and PostScript enabled rich design, but smothered the text’s semantic content in a welter of presentational commands. By removing presentation from the equation, HTML was much easier to author and kept its information content live – the feature that made search engines possible.
Berners-Lee deliberately ignored design concerns, but they soon raised their heads. In February 1993, Marc Andreessen, lead programmer of the development team for Mosaic, the first graphical web browser, proposed “a new optional HTML tag”, , that would enable GIFs and JPEGs to be embedded. Berners-Lee opposed the idea, preferring to finalise the original text-focused version of HTML and suggesting that a future extension of the existing anchor tag should handle graphics. Famously, Andreessen went ahead anyway, forcing the inclusion of in the first draft proposal for HTML, published in June 1993. By enabling HTML to handle “fancy graphics” as well as text, Andreessen and Mosaic can claim credit for turning the web into a full-blown publishing medium rather than a simple information-retrieval system. But by introducing presentation into HTML, Andreessen also opened Pandora’s box.
By 1994, his tone in the forums had changed. Replying to an inevitable query from a pioneering web designer about why their layout looked different under Lynx and Mosaic, his disillusion with what he now called the “ha ha, you can’t control what your documents look like in HTML” philosophy was obvious. To the suggestion that he therefore implement “one of the many stylesheet proposals that are on the table,” he replied: “Marc’s viewpoint: stylesheets are an artificial construct inflicted on us because of the whole non-presentation philosophy we’ve been using, which I argue is wholly inappropriate for document delivery front-ends and is crippling our system.” This front-end Andreessen was now working on was a rewrite of Mosaic that was to become Netscape Navigator, the browser, which captured the imaginations of both the general public and Bill Gates, and so turned the web into a global phenomenon.
Navigator addressed HTML’s presentational weakness through a slew of extensions that included new attributes to handle the appearance of existing tags and, more significantly, entirely new tags such as for controlling text size and face, and later various