Actually, HTML-based web application technology is obsolete. At least, it should be.

It’s time to start thinking about how to break back out of the browser. It’s time to re-imagine the Web as a source of data, not a repository of documents or a dynamic page layout system. A software interface is not a document, or vice versa. It’s time to separate the documents from the containers, and the containers from the applications. It’s time to use the right technology for each component of the system.

We’re getting close. Right now, the document is HTML, and the application is javascript, but the application is embedded in the document. Does that make any sense? The layout is the CSS, but the CSS describes the document, which is the structure of the HTML, but it’s also the structure of the application interface: the menus, the buttons, the text fields, the tables.

The application should be assembled not out of document elements, but out of user interface elements already present on the host operating system. Then the style can once again be focussed on the documents which the application displays. And other kinds of data, like tables and hierarchies and other heavily structured data, can be displayed using native user interface elements expressly designed for data of that sort.

A browser, in the Web sense, was created to display documents. Certainly, some documents contain more structured data, but a document is meant to be a static thing. An application is dynamic. An application is not a document, and a document is not an application. We have dedicated incredible resources to making them the same thing, now to the point of even subverting the continuity of the document to make it behave like an application, and at the same time we’ve made a mess where every application looks and feels different and alien compared to every other.

The web is a mess. Because whereas unlike desktop software, where there are rules and guidelines and established ways of doing things, there is chaos and disorder and countless re-inventions of the wheel, most of them inferior knockoffs and knockoffs of knockoffs. And most of them trying to use a document display technology as an application interface, and failing miserably.

The time has come (in fact, it came a long time ago, and its been waiting patiently for us to notice) to throw off the shackles of http and html and embedded web applications, and disentangle the streams of document data, document style, application data and application interface specification from the garbled spew which is the typical web (2.0) site. It’s time to write new data publishing protocols for the fast, efficient, and de-coupled delivery of highly structured binary and non-document textual data, and binary user interface description specification data. It’s time to put the code for user interface processing back into standardized, native code.

It’s time to treat documents like documents, data like data, and applications like applications.

The browser is a monstrosity. Let’s get over it.