1. Making the SL client multi-platform instead of Windows-only.

2. Releasing the SL client source code under GPL terms.

Some people believe that these were in fact Linden Lab’s worst decisions because they made content theft in SL much easier. But those people never really understood Philip’s vision: to create an open-ended 3D version of the World Wide Web.

In some ways, Philip was a Manchurian Candidate CEO. To venture capitalists he sold the idea of a virtual world with built-in protection of intellectual property and no way for users to download and save even their own creations, thereby locking them into the platform indefinitely. But many of his technical decisions were in fact undermining those plans. OpenGL instead of Direct3D? Quicktime instead of Windows Media? A client for Linux? Open Source? Don’t fool yourself into believing that he wasn’t aware of the consequences.

The consequence of those two decisions above is that Second Life as a platform will continue to exist even after Linden Lab (and their new competitor Blue Mars) is gone. The client/server protocol is open and fully documented, a Mono-based reference implementation of the server exists (OpenSim), anyone can set up a small version of the grid at home, connect it to other grids and then teleport between them as easily as browsing one website after another. Some of OpenSim’s features (e.g. megaregions) already go beyond the original server’s capabilities, without breaking interoperability.

I think Philip is leaving because he feels that he has reached his original goal. His job at Linden Lab is done. The genie is out of the bottle, and there’s no way to put it back there.