stating the obvious archives | about

The Veterans of a Thousand Browser Wars
Aug 25, 1998 :: Steven Champeon

Editor's note: As Netscape and Microsoft prepare the next major releases of their browsers, it's becoming clear to Web developers that the dream of cross-browser compatibility is quickly turning into a nightmare. While development cycles in Mountain View and Redmond are being spent on features that lock in customers, support for Web standards is being left by the wayside. As a result, Web developers are facing rising site development and maintenance costs due to differing implementations of the same "standards."

The Web Standards Project, founded by an astute and talented group of Web developers, has as its charter the education of developers, users, and the browser makers themselves as to the problems caused by browsers that don't fully support established Web standards.

This piece was originally posted to the Web Standards Project discussion list, and has been reprinted here (albeit in edited form) courtesy of Mr. Champeon.

Supporting standards is not about spanking vendors for introducing new innovations (read: proprietary and non-standard extensions) but rather about lobbying them to provide baseline support for the existing standards, so that those whose interests are best served by building to a common spec can do so, and those whose environments allow them the freedom to experiment and make use of non-standard extensions may also continue to do so.

We won't get there by arguing about which browser does it best -- the task is to educate and resist.

At issue are the skyrocketing costs of providing a top-notch Web presence, the entire groupware and directory services market, and the market for the desktop itself. The browsers are smoke and mirrors obscuring the real game. Netscape, by releasing cross-platform its original Navigator browser, turned the underlying OS into a commodity, easily discarded and replaced with others. Microsoft sees the future of its monopoly on the desktop as threatened by Netscape, Sun, and the Java-based network computer, among other challengers, and the only way it can win is to fragment the browser market, embrace and extend Java into the Win32 realm, and ensure that Windows NT lives through the next release cycle. Little of this has to do with the support or lack thereof for CSS1. XML, for example, is now being recognized as the driver that will open up proprietary file formats forever.

Look further. This is not about today.

We are all afraid that with the DOM spec being woefully unfinished, the ECMAScript standard lagging behind the implementations by at least two revisions, and the coming of XML, that the Web will become even more fragmented and even more useless than it is today. The vendors, whose executives and developers have seen standards wars before, need to be shown that the comparisons to the C or C++ or even IETF standards processes are empty. The stark, naked reality is that the languages governed by those standards were never seen by the end user -- they were compiled down to a platform-specific, but common object code, which is how your software is distributed today. However, in the space in which we circulate, it is that common object code, to use a metaphor, with which we deal.

HTML and the related standards (CSS, Javascript, the DOM, XML) are the "object code," the end result, and browsers must deal with them directly. There is no compilation step which might produce a different, but suitable, common object code. On the PC, the Intel monopoly provided the real standard for compilation, the Microsoft Win32/DOS monopoly provided the APIs to which source was written. On the Web, the standards are already "open," and a lot closer to the surface. When something doesn't work, it should be a lot easier to show your bosses and clients why.

Part of our problem as Web developers is that we are, to some extent anyway, a dime a dozen, the interchangeable cogs in a machine. Geocities just went public. How much closer would they have been to breaking even if the browser market wasn't fragmented? These are the sorts of burning questions which need to be asked, and publicly, by an informed and vocal majority.

If it costs $50,000 to develop a site that works across browser and degrades well, as opposed to $30,000, no Venture Capitalist is going to blink twice. But if it costs $5 million per year to maintain, and they see, or can be shown, that standards would edge them ever closer to the dream of every Internet company -- to break even -- then that is the course we must take.

We must strive to survive the browser wars, not prolong them.

-- Steven Champeon is cross-browser compatible, and is currently writing a book for IDG which discusses Dynamic HTML Graphical User Interface design and rapid application and prototype development. He is a consultant with, where he specializes in the design of large-scale intranet and Web site architectures and Web-based applications.



Other pieces about client-side software: