Behaviors
One of the major practical breakthroughs that resulted from the introduction of SGML was that it was possible to separate the appearance of text information from the information itself. For many professional technical writers (myself included), it meant that they no longer had to care whether they'd used the right font or not, or whether a product name should be in italic or bold. All the writer had to do was select the right element in which to wrap a certain piece of text and the authoring software would automatically choose the right appearance. Better still, it meant that the graphic designers responsible for the layout of documents could work independently, in parallel with the writers, without either party having to wait for the other.
Web pages have evolved in a similar fashion. Graphics and content came under the same roof until style sheets were developed, although many HTML pages still do not use style sheets. The appearance of style sheets meant that graphic designers and the content writers were able to divide the work between them. However, then people discovered interaction and animation; they wanted text that moved, and zoomed. Almost overnight a new bottleneck was created as everyone had to wait for the programmer to add the script, tweak the code, test the page, then hand it over for uploading to the server. Then there came the chaos associated with what script was best to use (Jscript, JavaScript, ActiveX, Java?) and where one found these script writers. Furthermore, you had to consider code re-use (the heart and soul of modern software development). Forget it-the closest you can get to that is good-old cut and paste!
It was only a matter of time before someone proposed a method of externalizing the behavior of HTML code-and both Netscape and Microsoft have tables their own solutions.