It all starts here in one form or another: HTML

• What started as a language defined by SGML (Standard Generalized Markup Language), HTML is used to mark up billions of pages, making up the bulk of the web.
• From a common origin, 3 distinct version have emerged:
  1. HTML 1 - 4.01: This was developed in rapid succession in the 1990's
  2. XHTML 1, 1.1, 2: A stricter version, created as an answer to the loose development of HTML
  3. HTML5 - a reaction to the draconian ruleset proposed by the W3C for the future of MLs beyond XHTML 1.1

• Late 1991, Tim Berners-Lee releases the first iteration of what would become HTML 2.0 (there was no real 1.0+)
• HTML was a language created at the dawn of the Web, using SGML (Standard Generalized Markup Language) as a template
• Over the next decade, various additions, improvements and changes are introduced
• The last official specification was the service release of HTML 4.01 in December 1999
• Initially used for documents and a place for the technically inclined, the web did not have the international pervasiveness it does today
• Most developers were new to the language and there were no real classes in it
• People had to learn as they went along
• A lot of early sites are poorly designed

HTML 4.01 template

• Despite it's reputation as a loose language, HTML 4.01 can be written strict just like XHTML and HTML5 can
• The problem was not entirely with the language or the authors. It was with the browsers.
• Browser manufacturers created User Agents that would read code, interpret it, fix any mistakes and render a page, without any visible errors.
• It was this that led to coders exploiting problems in browsers and creating pages that broke when new specs or updated User Agents (browsers) came out
• Okay, so now there is a problem. Millions of pages are being produced with sloppy code because the UAs let it happen
• The web is becoming filled with Tag Soup that makes it hard to index, with search tools having to trust things like META keywords and descriptions
• A better solution is required: Get Strict!

Moving on to XHTML1

• After years of HTML development, 4.01 became the "Last Official Release"
• The Web Working Group had taken a look at XML (Extensible Markup Language) and liked what they saw: a requirement for valid markup in order for an XML application to work
• Why not take the ruleset from XML (lower case attributes, everything in quotes, close every tag) and apply it to HTML? Yay!
• That should fix everything, right?
• Coding wasn't actually fixed by the introduction of XHTML. Fixes were suggested by the specification, but most browsers would render the page anyway.
• This is because 99% of XHTML pages are reporting their MIME type as: text/html
• To get pages to render using the appropriate DOCTYPE of XHTML, pages would have to be delivered with the MIME type of application/xhtml+xml
• Using that MIME type, however, brought with it a rather dire proviso:
• The error handling specified in XHTML 1.1 meant all errors will cause the browser to cease rendering
• Yes, you read that correctly. That means any error will throw an warning to the user and the page will not be shown
• Try to imagine the Web actually working in this way
• Why was this done? Because the developers of XHTML wanted to force strict coding rules so that the Web could be seen as a giant application. Reliable code on every page meant you could use the data anywhere