Can Markup Unite?

Getting past the HTML / XML splits

A few years ago, I stopped talking about XML and starting talking about markup. After a few too many conversations with developers who had found XHTML, web services, and various other things that had proudly branded themselves with the “X,” it was clear from hostile responses that XML’s boom was done.

Many of Wednesday’s Balisage talks wrestled with the challenges XML tools face in a world dominated by competitors, especially HTML5. Today opened with a talk on making XForms work in HTML5 with browsers, followed immediately by a talk on replacing DocBook XML with XHTML5 (at O’Reilly). Two more abstract talks looked at filtering and an info space model, but the afternoon asked “Where did all the document kids go?” and then sought world domination by making XML invisible.

Line at Balisage

While it was a great selection of papers, there was some angst around the basic problem that XML-centric tools are having a hard time competing with HTML5 and other formats.

  • The XForms talk was a look at implementing a set of tools with a dedicated community that the browser vendors sidelined in favor of HTML5. I suspect that some kind of polyfill approach could work well for it, letting it run happily in an HTML5 context despite the challenges.
  • O’Reilly’s shift from DocBook XML at the center of our publishing universe to HTML5, even XHTML5 with book-specific extensions, was not entirely popular with folks who had long seen DocBook as a flagship of XML, or with people who considered HTML semantically impoverished.
  • The comparison of XML tool history with the history of other web and database tools made it sadly clear that XML tool building, at least on the open source side, has lost its momentum badly. Many, many people (including me) are using older libraries like LibXML, a great tool that has been largely in maintenance mode for a long while. I asked whether there was enough community left to restart the tools chain, or if it was time to rebrand and reconsider.
  • Steven Pemberton revved up the crowd by having us plot and celebrate XML world domination, but Invisible XML is a great way to treat non-XML content as XML. Folks who want to treat information as XML will gain a lot, but XML itself, as opposed to XML data models, won’t gain much traction this way. If anything, this model is a (frequently appropriate) way to use XML itself less.

I’m not sure that my personal rebranding of XML, HTML, etc. as “markup” resolves these genuine tensions. Markup changed the world through the Web and to a lesser extent through XML, but remains an often misunderstood corner. I regularly have conversations with people who think that an HTML document should just be a header that loads code to build an object model, and the idea of humans editing markup seems to bother some programmers’ clean room fantasies.

I hope that view from the outside could shift if various markup communities can come together eventually and define a less-fragmented practice. Unfortunately we don’t all know each other that well or trust each other that much. Is this a good time to work on that?

(The Balisage conference remains the place to go if you want to think about how humans and computers interact with markup. I sometimes refer to it as “the last standing XML conference in North America,” but the reality is that it’s more than XML and lingering memories of its SGML parent: other markup is also welcome. Unlike most conferences I attend, it has proceedings, with papers for most sessions. With six years of papers, you could be reading for a long time.)

Related

Sign up for the O'Reilly Programming Newsletter to get weekly insight from industry insiders.
topic: Web Platform