Randomize

Note: this was originally published on 2003-07-02. The predictions came true, but not along the path I expected. Instead, we’ve seen rich JavaScript-based UI frameworks like ExtJS make “native-ish” web application development a snap, and Flash and Silverlight (the latter being closest to my expectations) have become the de facto, though controversial, replacement platforms in the browser. Posted here for posterity. If anyone has a way to easily convert old dasBlog content to WordPress, please let me know…

First, Microsoft gave IE the “kiss of death by atrophy,” promising nothing new until Longhorn. Hints fly that Microsoft is going to reinvent the browser, bringing it to the next level (I presume they are not talking about standards compliance here).

Meanwhile, the Mozilla roadmap shifts focus away from a “mega-browser” and toward Phoenix–light, fast, and complete with its own fully-skinnable forms language, XUL.

Now, Marc Andreessen (co-author of NCSA Mosaic, the first web browser) is complaining about lack of innovation in browser technology, specifically harping on navigation limitations.

Related? Yes. Conspiracy? No. Just the sound of Necessity mothering multiple inventions. Here’s my predictions about the next evolutionary step in the world of browsers:

• (IE) WinForms.NET support hosted directly in the browser, integrated with scripting languages. Fixed layout will be possible, so form elements, though on the canvas, become part of the application chrome without hacks like .
• (Mozilla) XUL does the same thing. Someone will write a translation layer between WinForms and XUL. Mozilla still has two advantages: skins and speed.
• (IE) Browser-side Javascript and VBScript are augmented with the browser natively hosting any CLR-targeted language, with full DOM access. This is not the same thing as current .NET methods for downloading and running WinForms apps from the browser.
• (Mozilla) Mono is tacked on to provide similar functionality.
• (Someone) Bookmarks/favorites replaced with a database that automatically sorts, files (in multiple categories), and retrieves links, offers related links, handles broken links, and adds location independence through web services (like the wonderful, transparent bookmark synchronization software/service offered by SyncIt.

In the end, we get the best of both interface worlds: browser-like flexibility with application-like usability. More advantages:

• Easy integration of rich content, “scrolling” forms and reports, images, and multimedia elements into any application, with all of the layout flexibility of XHTML.
• Non-scrolling, site-specific menus and forms rather than quirky HTML menus, buggy Java appplets, and limited HTML forms.
• Native form widgets also means user-theme-sensitivity and accessibility.
• CLR speed, features, type-safety, and language-independence for browser scripting.
• I can finally have a real form button (i.e., akin to <button/> or <input type=”submit” />) with a transparent PNG icon and label that actually looks right.
• Any web site can expose an interface that is indistinguishable from a native app (think Microsoft Money for an example, it makes a *lot* of use of MSIE’s renderer).
• Any application can expose rich content as simple as spitting out some XHTML, XML, SVG, etc. No need to reinvent the layout and presentation wheel.
• Web services finally find a good use as a real replacement for traditional HTTP POST/GET. Currently, they are only generally used for server-to-server or rich client to server applications.
• Event-driven forms actually begin to make sense now that web applications are not limited to the page-stateless model. Event handlers can be created on the client or server or some combination via web services.
• Scripting on the web browser side can be strongly encrypted to reduce chances of tampering or bypassing by the user. Data channel encryption not limited to HTTPS.
• The experimentations in integrating IE and Explorer for a richer interface reach a higher level of maturity, and Outlook and Outlook for Web merge into a nearly-identical codebase.

So, Robert, am I anywhere close? [Scoble mutters something about NDAs and ice cream and how “Google still works…” 😉 ].