Obsolete websites and Sideways compatibilityThursday, September 12, 2002
The Slashdot article has highlighted a false perception of what typifies "other browser". Quite a significant portion seem to believe "other browsers" means older browsers identifying themselves as Netscape Navigator and Internet Explorer. It looks like they completely overlook the real _other_ browsers out there, the speech readers, the line-mode browsers, the scripts out there that try to do useful things with web content.
For as long as the website designers ignore the arena of "non-visual" user-agent, they are not going to see more than the "two-browser world". The web has far more depth than that. From one of my Slashdot replies:
Its not the backwards compatability that concerns me, its the _sideways_ compatability that's more important to me. The authored HTML tends to work in a range of Netscape browsers, a range of Internet Explorer browsers, and sometimes in a range of Opera browsers. Anything other than that is random.
A standard's adhering HTML document could be used in all the browsers above, plus all the other user agents out there that support the standard followed. So text-to-speech browsers, indexers, spiders, content aggregators -- all the silent user-agents suddenly have access to structured content.
These are the user agents that are overlooked by the typical public website. People don't tend to notice that structured markup scores a lot better in google than font-flavoured tag soup, precisely because h1 defines a first level header, and font defined some weird presentational style but nothing semantic that a search engine can use.
I don't believe browsers will be the user-agent of choice in the coming years - we'll automate all the manual intensive process of trawling through websites looking for information, and we'll delegate it to some sort of intelligent agents that do the work while we do something more enjoyable.
RSS Aggregators like AmphetaDesk show a very basic inkling of what can be possible with structure and the value of content out there on the Internet.
But we need structured markup to add semantic meaning to the content, and then we can leverage that content into something truely useful. (Yes, I'm a dreamer longing for something practical)