Bonum Certa Men Certa

Microsoft on the dangers of a web-centric desktop

Date: 1998



From: Darryl Rubin Sent: Thursday, October 22. 1998 5:46 PM To: Scott Cottrille Subject: RE: Annotations


Deeper Web/Shell/App Integration: The HyperActive Desktop



In this memo I want to talk about extending the efforts we’ve made for unifying the shell and browser and suggest how We can take these ideas further; how we can:

¯ Further integrate apps and web browsing into the shell frame to achieve a navigation/viewing experience that is both richer and more seamless

¯ Better exploit push model to reduce information overload

¯ Simplify information access and filing by support of unified storage

¯ Use rich linking to make collaboration and sharing inl~erent and convenient features of the shell.

A couple trends motivate these ideas. One is that the PC is becoming a device more for communicating, collaborating on and consuming content than it is a device just for producing it. The other trend is that the web is becoming the primary metaphor for these consumption and collaboration activities.

The threat to Microsoft is that companies like Netscape and Lotus will be able to offer web-centric "desktops" that users will prefer living in and using as their launch points because they better support the user’s web-centric metaphor.

This is not hard for our competitors to do, because the amount of local desktop functionality they need to subsume isn’t that great (the Windows desktop isn’t super rich, especially in the consumption/collaboration departments) and it’s easy to design a set of HTML pages and controls that put a pretty nice face on the existing local desktop operations. This is, after all, partly what we are doing in Memphis.

The danger of competitive web-centric desktops is that they will be sold as platform-independent shells, shells that have an especial affinity for Java apps over native ones. This threatens the core of our platform and application strengths.

The challenge for us, then, is to enrich the desktop in web-oriented ways so that it is not so easy to replace, and so that running Windows-native apps under the Windows-native shell---especially Office--results in compelling advantages compared to Java apps on a platform-neutral shell. The web/shell integration work we’re doing in Memphis is a good first step, but there is much more we can do.

Full Exhibit