Does Linden still plan to open-source the server?

[Mircea Lobo]

More than the Linden grid, I'm interested in using Second Life as an open and complete technology. Although I've been with the OpenSim project for years, there are still a lot of things that are not ready yet in OS. I was googling about the Second Life server, and found old Linden announcements saying the server side might be open sourced as well.

At this day, is there any information about Linden's take on this? Do they still plan to release the original server code like they did with the client, and is the matter being discussed? It would be a nice thing and highly appreciated if they could do this at some point.

Note that I'm not suggesting a way that would allow stealing assets or other SL works... but for using on an entirely different grid (or hosting your own sim on the LL grid but without being able to copybot things or hack objects).

[Ceera Murakami]

Linden Lab has never intended to open-source the server code. Only the client.Some users speculated about the server code also going open source, but LL canned that idea almost as soon as it was first mentioned.

If the OpenSim grids could offer 100% compatability with the main grid's functionality, it would hurt SL a lot more than their half-implemented, reverse-engineered server code does today.

[Mircea Lobo]

Interesting. I wonder how articles like this one or this one appeared 5 years ago. I was hoping they're part of a delayed decision or something. Still, is LL determined not to consider this at some point? As for harming SL, a fully compatible server might affect the main grid, though I don't think it would decrease its popularity or user count too much. Not to be mean, but I still wish a lot for such a server to happen and offer us an alternative

[Ceera Murakami]

Joe Miller may have made statements like that 5 years back, but I think he was talking out of his own belief, and not stating formal company policy. Or that if it was a "plan" at the time, the response the announcement generated caused them to do an immediate about-face on the idea. Certainly they never did release the server code to open source. Intsead, they tried licensing a turnkey server system (SL Enterprise), which tanked horribly.

Is Joe even still around? Or did he get kicked out with one of the employee purges?

[Sayrah Parx]

Philip Rosedale also stated in this interview from 2007 that they were already making preparations for it to become open source.  However, they eventually decided to try making more money from businesses and institutions with licensed servers that could operate within the privacy and security of an intranet.  Making the server code open source seemed to defeat that plan from their perspective, although the success of things like Red Hat Enterprise Linux should have suggested otherwise.

They obviously need to start over from scratch and build something more scalable, based on the lessons they have learned over the past decade, so I'm not sure why anyone would want the current code anyway.  A dedicated start-up team would be in a better position if they never saw any of the Linden Lab code, if only for the sake of their own ingenuity.

The major failure of OpenSim is that they depend on compatibility with viewers based on the open source Linden Lab viewers, which rely on the fundamentally unscalable Linden Lab servers.  The server can only be as good as the client, and vice-versa.  At Linden Lab, a big part of the problem is that the server and viewer teams are completely different people.  For a new grid to be successful, they would need to develop their own viewer and have every developer working on both the server and client.