The World Wide Web

There are only two things wrong with the World Wide Web:

  1. The mark-up language
  2. The protocol.
Other than that it's fine.

The Mark-up Language

Its initial aim was to make scientific papers available to a world- wide readership. At the time that the World Wide Web was being developed at CERN, most scientific papers were written in a quite different mark-up language called LaTeX. LaTeX has excellent support for equations. Support for hypertext links could have been added to LaTeX, a web browser could have been written which displayed LaTeX documents, and all those scientific papers could have been put on servers without rewriting, providing the web with its initial content.

But it didn't happen. For some reason which baffles me, HTML, a variant of SGML, was chosen instead, which didn't handle equations. There were no documents written in HTML before the WWW appeared, and no HTML viewer until Mosaic was written. To publish those scientific papers on the web, it was necessary to convert them to HTML. There are tools to do this, but the end result doesn't look as nice and equations can't be handled other than converting them into images, which cannot be edited.

So while what we have is a pornographer's dream come true, the boffins back at CERN, the World Wide Web's birthplace, still are inconvenienced when publishing their papers on particle physics on the web.

Then, because no one knew HTML, many web pages were written in malformed HTML. Web browser were developed to handle this. Many HTML documents were generated from Microsoft Word documents using Word, which (I suspect deliberately) output malformed HTML. The other browsers then handled this. (Microsoft has now cleaned up its act and now balances tags, though the HTML generated is, like all other things Microsoft, horribly bloated.)

The Protocol

Who controls what is displayed on your browser? The webmasters, that's who. Wouldn't it be better if you did?

A long time ago, about 30 years before the WWW appeared, another hypertext project (Xanadu) was started by Ted Nelson and only recently began to bear fruit (Udanax). Xanadu would have allowed for backlinks, annotation of other people's web pages, the ability to quote excerpts from other people's web pages (with a link back to the original, and even micropayments. The problem is not so much that HTTP doesn't handle any of that. It's that HTTP can't even be extended to handle it. We're stuck with things as they are for the foreseeable future. Yet again we have seen Worse is Better defeat The Right Thing. I guess Ted Nelson's Belbin type isn't Completer-Finisher.


This page was linked to from

and was last updated on Feb 28 at 23:19.

© Copyright Donald Fisk 2004