3pointD.com

I received a nice note from Cory Ondrejka, chief technology officer at Linden Lab, makers of the virtual world of Second Life, in response to my looong post of Saturday entitled “Linden Lab Approaches A Crossroads.” I won’t quote from it, as it’s not an official LL response, but I’ll raise a couple of the points Cory mentions in order to give a differing viewpoint on a couple of things.

Cory notes that the Lab now has an opportunity to build a standards-based architecture that would allow SL to scale similarly to the World Wide Web. I’m all for it. I’m not at all against the effort to re-architect the Grid; I’m aware that it’s been under way for some time, in some form. It’s good to see that the Lab is pressing forward.

Cory also notes that new hires will allow LL to expand its operational team to cover more time zones, which will help it recover from problems that occur when California-based employees are happily asleep. That can’t be a bad thing either.

I had complained in the previous post that locking out potential customers was possibly bad business. Cory responds that, personally, he’d rather not expose them to a laggy grid. I can see his point, but I still think there’s a danger that they won’t come back for a second try. There’s clearly a danger either way, and it’s hard to tell which is greater.

One or two points that I might need to clarify from the OP:

When I wrapped up by saying, “It’s no use striking off toward three different futures at once; you’ll never reach any of them,” I wasn’t referring to different directions in engineering efforts. Several people have noted, and I agree (and noted in the OP) that the Web team has a different skill-set from the server architecture team. I was referring more to the timing of the Web-based effort, and to higher-level policy decisions to take on new tasks like LL’s building its own Web services, rather than simply opening a broader API for users to build on top of. The decision to limit access for unverified users also raises questions, for me at least, of how open a platform SL is intended to be. In the end, though, I have a lot of faith that it will continue to be an exciting place no matter what decisions are taken at the Lab. (Well, there are probably a few choices that would make it a lot less exciting, of course.)

My bet is that the changes will probably improve the Second Life experience, eventually. But I still feel like the Lab is going in too many different directions at once. The system of servers, clients, and asset- and information-management systems that underlies Second Life is not quite stable, and not yet even in final form. Re-architecting the Grid is fine and necessary, but when you’re building other services on top of a system that’s in the middle of being transformed, there’s too much risk that you’ll see a lot of that secondary effort go to waste. I’ll be the first to note that it’s easy for me to armchair quarterback the whole thing, but if I were an outside developer (and these now include the likes of IBM as well as a raft of individual coders, and everyone in between) I’d feel more at ease if I knew that the Lab was going to lock down a Grid that would last before they set about building services on top of it.

That said, it may be the case that the architecture of SL’s 2D information systems — account information, friends lists, etc. — have already been locked down to the Lab’s satisfaction. But if so, why not provide an API? The two very limited APIs that the Lab does provide are made available only to certain types of users under certain conditions. (A Web site that runs an L$ exchange, for instance, can find out what risk LL assigns to my account, but I cannot.) This is hardly the kind of Web 2.0-style openness the Lab professes to be about. In fact, it hews very closely to kind of gentle discrimination implied by the locking out of unverified users: the Lab continues to make judgment calls about who may use different parts of the service, at which times, and for what purposes. If Second Life is to grow to the scope that Rosedale envisions for it and toward which its engineers are working so hard, that kind of judgment call will have to stop being made.

LL will have to proceed carefully through this current phase of its evolution — more carefully than it proceeded in its early stages. The ramifications of the design choices that were made then are only now being felt and corrected for. The Lab is of course in the unenviable position of attempting to build something no one else has ever built before — a 3D online world that can support unlimited users and which features open content creation tools — but that’s only one more argument for slow and careful attention to the Grid’s genetic code. (My perception is that the process of mutation is moving slightly too fast.) It is conceivable that they’ve set themselves a task that can’t actually be accomplished. But that has yet to be determined, and you have to applaud them nonetheless for pushing the metaverse as far into the future as they have, and for even trying.