Nigel McFarlane

Cleaving the Body Politic

Written by admin on May 5, 2009

Whether the Web is defined as HTML, HTTP or the dynamic and discriminatory mouse-clicking of Web surfers, there are larger perspectives of the services that Web technology enables. We can use these larger perspectives to qualify and test how some of this rough treatment should be assessed.

At the simplest level, the Web can be seen as the exchange of documents cast in agreed formats. Document formats are produced and consumed by machines, except in the near-mythical case of “plain text”. Where no users are directly involved, it is clear that there is no excuse for failing to implement document formats – machines are quite good at rigidly specified tasks. Thus, mistreatment of HTTP headers, which are machine-crafted in all but the most extreme of environments, is justifiably viewed as an inexcusable piece of rough handling. So much for Internet Explorer. Attempting to apply this logic to HTML must fail, however, because numerous HTML documents, including this one, are still hand crafted. That means that HTML is not yet exclusively a file format.

A more expansive view of the Web is that it is a global distributed application for document management (DM), either in terms of technology or in terms of user behaviour or psychology. If this view is taken, then it follows that the built Web is a very rudimentary document manager indeed, since standards exist at the most basic levels of DM functionality only. Even RDF and OWL, the most abstract of the Web standards, say nothing about the common structures required for the numerous vertical niches within the document management field, or anything about basic change control. The document management argument is thus a “no-case-against” argument for rough handling of standards. Web browsers and servers are sufficiently primitive at document management that any advancement is water to a thirsty man. There is, from the document management point of view, no case against DHTML Web application hackery, no case against Web surfers accruing bookmark collections as large as an encyclopaedia, and no case against Web browsers being hacked, optioned and special-cased to within an inch of their lives. Such things are all required when underlying standards are so grossly inadequate from an application point of view.

There is a historic Microsoft analogy for this set of circumstances. One could not expect a modern, graphical desktop from MS-DOS: at best one could expect stop-gap tools like Sidekick and XTreePro. MS-DOS was just too primitive alone to be a full solution. The layering of Microsoft Windows 3.1 and other equally aggressive applications on top of MS-DOS, or the abandonment of DOS entirely in favour of the Apple Macintosh were the inevitable consequences. It is not unlikely that such a process will repeat itself in the case of the foetal document management functions that the currently built Web represents.

Less bleakly, some prefer to see the Web in terms of discourse. In this paradigm, the built Web is a lingua franca – a common language. The logic runs that the electronic discourse that is the Web, as facilitated by HTML, is somehow a pillar of civilisation and requires respect equal to that accorded to the courts and to football players. People must be able to communicate with each other – otherwise fragmentation, misunderstanding, woe and eventually war will be the only consequence. If there is not some minimum standard that HTML browsers require, and if Web page creators are not assisted to reach those minimum standards, then trains will no longer run on time.

There are several problems with this sort of logic, which hopes to prevent the propagation of poorly-formed HTML and perhaps also hopes to bring about the death of the MARQUEE tag and its raggedy-bottom associates.

The easiest of these problems is the supposition that a lingua franca is required. It is not. The Web is not Northern Ireland or any other equally torn state. No one is born into strife on the Web. Use of the Web is an example of joining an intentional community, whether that joining be as a supporter, an opponent, or merely as a garrulous ratbag. The conditions for intentional communities are well established: if you don’t like it, leave; if everyone doesn’t like it, the community falls apart. There is no need for a de jure common tongue in such places, because the surfer enters into precisely the dialog desired, with whomever it is desireable to dialog with. The Web can be arbitrarily partitioned into as many such intentional communities as you might care to identify, each of which might have their own perspective on HTML. In any event, the clashes of ideals and flame wars that occur from time to time on the Internet are testimony enough that a few HTML tags are not enough to ensure civilised (or normative) behaviour of any kind. So civilisation building, even of electronic civilisations, is not sufficient to support raising standards of HTML use.

A far more serious problem with the argument for quality-HTML-as-discourse revolves around markup. The markup concept is more than sufficient reason to ensure that rough treatment of HTML will and should continue for a long time.