It's a fine mess we're in

Soundbytes on the World Wide Web

The days of the World Wide Web are numbered. No one can deny that it transformed the internet from a vaguely interesting academic computer network, with some commercial potential, into the core infrastructure that it is today. But it is now time to move on.

Back in 1994, companies were buying internet access from the early commercial providers such as Pipex and Demon and wondering what to do next. E-mail was useful, but copying files was dull and complicated. And while reading Usenet bulletin boards gave systems administrators a way to pass the time, there was little else to do with the expensively acquired link to the "information superhighway". The web - or rather, the Netscape browser - changed this by offering internet users who lacked computer science degrees, or an interest in obscure programming languages, an easy way to find and read documents of all types.

Yet after ten years of worldwide domination, it is time we accepted that the limitations of the core web technologies make it unsuitable for many of the niches it currently occupies. We have to move on, and let the web take its place in history instead of allowing web-centred thinking to dominate the online ecosystem.

Visit any commercial website and use the "view source" command to look at the raw code of the page - the HTML - and you can see the mess we're in. You are likely to find dozens of lines of code put there to figure out which browser is in use, so that the "correct" version of the page can be sent over. Clearly, something has gone wrong.

The early web developers aren't to blame. Sir Tim Berners-Lee, the English physicist who invented the web while working at the Cern laboratory back in the late 1980s, was trying to link published research papers to each other. He wasn't planning the global information infrastructure the web turned out to be, so it is not surprising that the technical choices he made no longer serve us well.

A website today has to be interactive, e-commerce enabled, customised and, above all, accessible. You can modify what already exists, just as you can convert a shop to give wheelchair access, but too often the result doesn't satisfy anyone.

Some companies believe that building accessible sites is so complicated that it will vastly increase the cost of any development, thereby rendering the site a commercial failure. These are the people who gave us the original Odeon cinemas site, the online equivalent of an office building with steps to every room, 50cm-wide doors, and signage in an obscure Tahitian dialect.

Good design can produce sites that are both functional and accessible, but doing anything visually interesting on a web page does reduce its accessibility. That is why Tesco decided that it needed two sites; one for the typical user, and one built to accessibility guidelines. However, it could have done something more innovative and built a shopping program that did not use a conventional browser, but linked the shopping application directly to the screen reader or other accessibility tool. The web mindset is so dominant that it didn't even consider it. E-mail, of course, has always been distinct, but instant messaging programs are also installed separately from the web browser, as are the programs that allow millions to benefit from free online phone calls with Skype.

We have tried having all our online services provided through a single client, the browser, linked together in a way that makes it difficult to see what is connected to what, but it just doesn't work. We should have a wealth of collaborating programs, designed to work together, based around sensible user interface standards that are able to deliver the services we need. Otherwise, we run the risk that the internet will stumble, fall, and eventually die away.

Bill Thompson writes a regular technology column for BBC News Online