Web Pidgin

By Paul Ford

Standards and schemas, pidgins and islands.

I spend hours each week reading dry technical documents to learn more about my trade. I can reel off about 30 Web standards, all of them filled with rules for the right way to build a site, all of them at lea st a little confusing and difficult to read.

Why all these standards? In the 90s, an awful lot of people set sail to this island called the Internet. Some of us came from TV, or publishing, some from newspapers, or the art world. Some, like me, just got out of college and needed a place to go.

So we arrived on Internet Island from these different countries, and when we landed, we found that we couldn't understand anything said by anyone else. Television people wanted the Web to look like TV, with lots of video. Print people wanted good typography. Marketing people wanted to target their ads and count the hits. The software folks thought we were all crazy interlopers.

As happens when cultures bash into each other, we created a pidgin language of standards, acronyms, source code, and tags. Take XML. XML was supposed to replace HTML by simplifying SGML. Well, come on, sure, who doesn't want that? XML was going to add semantics to data so you could make smart searches, you could say, show me all the poems about red wheelbarrows, and boom, there you'd go. Except that didn't quite work out, so a few years later they created RDF, which allows you to be much more specific about what you meant by poem, or red, or wheelbarrow. But RDF needs something called RDFS. And that was so complicated that they had to create OWL, O.W.L., to clarify things. Which would be great, except you need to understand everything all the way from OWL down to XML to get any real work done. Which is why some pidgin-speaking Web developers drink in the afternoons.

Over time, pidgins become creoles. A creole is a full-fledged language, a mix-up of all the other languages that went into it. But you can't force a creole into being. It has to evolve. Most standards would like to be seen as creoles, as natural progressions from the earlier pidgins, but they're more like Esperanto-an invented language that you have to learn from scratch, defined only in terms of itself. What web developers actually use is not pidgin or Esperanto, but pidgeranto, or esperidgin, a kind of pidginized Esperanto. Because at some point your clients just want it to work. There's no time to study any more standards. You need to publish the site. So you make do with what you've got.

We've got a long way to go before the Web has its own native tongue, one that isn't simply a quilt of style sheets and XML and JavaScript, but something unified that you could speak fluently if you tried. It would be great if you could force the process, bring that language into being, but you can't; it has to evolve over years, over thousands of interactions and discussions and implementations. So in the meantime, we speak pidgeranto, which is much better than the pure pidgin of a few years ago, and the standards stack up, each one a step towards something complete, something whole.