Innovation, yes, but wrong turn

The way Linden Lab is also really considering to outsource some of its work to in-world communities shows how Rod is slowly steering LL back to their pre-M Linden days. In fact, to its “very early Philip Linden days” — the days LL seriously encouraged projects, from Mentors to self-sustainable communities, and which were all abandoned under the umbrella concept of “favouritism”. I don’t know how LL is going to deal with that ugly word nowadays. Probably they will just wisely ignore it, and really go ahead with more tight collaboration with the residents willing to work with them. I, for one, am all for it: it’s a vision that I wholly subscribe to. I’d certainly would like to have some kind of governance with resident participation, but, alas, that might be too early: Rod is testing the waters first and not taking huge steps. But the ones they are taking have been encouraging so far; if anything, the only complain at this time is about how much LL is really going to allow residents to get involved in “our” virtual world again. Will it be “Your World, Your Imagination” again? It’s early to see, but it looks like it. I’m crossing my fingers

So what are people really complaining about?

OpenSim-based grids are a viable alternative these days, even though not all can provide the same kind of robustness that LL can. The reason is simple: the OpenSim software has been designed mostly with the goal of replicating LL’s functionality and allow it to be extended on an experimental base. If anything, it’s “research-quality” software: excellent to tweak and change, fantastic to play around with, and always surprising to see how well it performs in controlled conditions, but it’s not “industry-grade” — not yet, even though a lot of operators claim the contrary. They can very well claim that, since, on average, their grids are just a tiny fraction of LL’s own. The increasing complexity wreaks havoc with the infrastructure model, and OpenSim, while ever-so-resourceful, is still not at the stage that it can seaamlessly provide service on 30,000+ regions with 50-100,000 simultaneous users.

But the point is that it doesn’t have to. It can work as a standalone grid or as a federation of grids, interconnected via HyperGrid (although as an operator grows, the tendency is to shut HyperGrid down); this means that in theory at least, OpenSim grid operators can “keep it small” — and operate large numbers of interconnected grids, but each one being relatively small in size, thus effectively dealing in a different way with scalability problems that Linden Lab does with its own Second Life Grid. The major difference is that some products and services can only cross grids with some difficulty. Nevertheless, we have now money exchanges that work across grids, and we also have web-based shops that can deliver to different grids. Whatever OpenSim’s future is going to be, it already shows that it might very well be a large, federated metaverse.

In the mean time, it provides an opportunity to contrast it to Second Life. And what operators have found out is that the biggest complaint is really how much it costs to be a resident owning land in SL.

Now I’m not going to do the calculations again to show that LL is not really “ripping us off”. Every year or so I do the calculations again and post it as a comment to some blog, and then immediately forget where I wrote it The short of it is that LL is profitable, yes, and will continue to be so even with a slightly declining landmass, but they’re not really making huge profits. The overall price structure for Second Life is actually adequately indexed to the costs of operation.

So how can OpenSim grid operators actually deliver similar (not the same!) service for a fraction of the cost — sometimes as little as a tenth of the price, sometimes even for free? The answer is very simple. It’s not that OpenSim is incredibly more efficient and can be deployed at extremely high densities (of number of simulators per server). No, the simple reason is that OpenSim grids have absolutely no software costs. Linden Lab, by contrast, has to maintain several teams just for the viewer, and another set for the four simulator “channels”, plus a host of web developers which work on the “glue” between the grid and some services (like the full range of web servers operated by them). Plus, of course, technical support, an engineering team, and all the associated operations staff. This ads a lot to the service they run, and I’m not even considering the administrative costs. While OpenSim grid operators usually can get away with one network/infrastructure/system administrator, one programmer (to provide “glue” as well), perhaps a content designer or two, all of which may double as technical support, and little else. Even though Linden Lab benefits from reduced costs for having so much infrastructure — there are huge savings due to scale — they cannot compete with a tiny operation running a few hundred regions with perhaps a staff of a dozen people, most of which might not even require office space.

Most people, even seasoned web developers, just figure out how much it costs to rent a good server on a high-end facility, and knowing that such a server can host a lot of sims (well, at least one per CPU), they figure out that LL is “robbing” with their high prices. Well, they forget the huge staff that LL actually has to keep.

Nevertheless, even if some residents “understand” business costs (few actually do), they still compare prices. With such a huge price difference, and linear growth stagnating (which is bad for the ones still believing that the 2005-2007 kind of exponential growth is still possible somehow; more on that in a moment), it’s obvious that residents start asking why they are spending so much, and what real benefit they’re getting from their investment in SL. I think this is a natural question to ask. And the answer is actually honest: the cost is paying for a lot of privileges. But the problem is that Second Life’s economy is not up to the expectations of many land owners. They miss the days of exponential growth. It’s not really about “what do I get for US$295/month”, even though that’s the kind of questions that get asked. It’s a bit more convoluted: “Back in 2007, I would pay US$195/month and get cartloads of residents on my location, shopping either for land or goods. This has all stopped. Now I’m forced to pay US$295/month for next-to-zero visitors and almost no sales. Why?” And then we get people blaming Linden Lab for the lack of new users — lots of new users!

Pan to the next scene. Linden Lab still believes that somehow, by some trick, they can tap again into that powerful exponential curve again. They figure out that everyone who was not impressed with SL in 2007 will have newer, faster computers. There is way more high-quality content these days, too. The grid is far, far more stable. Even group chat works again! So, having solved most of the more pressing technological issues, and adding a lot more features, a new viewer, and so forth, they guess that this should be enough to drive more users in.

And guess what? The number of new registrations per day almost doubled with the “light” version of SL 2.X being available, plus a more streamlined sign-on process, with a better orientation area. This was exciting news for LL. Maybe SL can reach the mainstream after all? A few tweaks here and there, zero advertising, and the gates seem to be flooded with newbies again.

But no. All these people just sign up, log in for a few minutes, and never try SL again. Something clearly disappoints them. Puzzled, Rod Humble announced a certain degree of gamification. Rod’s background is in game design, so it’s obvious that he knows what that’s all about. He reasons that if SL’s initial experience looks “more like a game”, gamers will be more familiar with it from the start, and might not log off immediately, but stay around for a bit and see what’s all about. The new sliding panel on the Basic Mode, which shows destinations, and helps new users to pick an avatar very easily, is also a very easy way to get people familiar with SL — just point and click, and you’re going to see fantastic things you never dreamed that were possible! To make sure that lag is not an issue, I believe that Basic Mode will, by default, run at the lowest possible settings for the graphics card. And as I’ve come to notice, these days, even the lowest settings are impressively good. Being cursed with a almost-6-year-old-iMac, which doesn’t even support anything beyond OpenGL 2.0, I’ve been delighted to see SL running steadily at 30 FPS at next-to-lowest settings, but which look so extraordinarily good on SL 3.0.3, that I don’t mind at all (sure, for taking pictures, I might throw in a shader or two, but, in general, why bother? SL looks hot even on the lowest settings!)

However, I still think that LL (and not only Rod; he’s just starting to think like the rest) is working from a totally wrong assumption: that virtual worlds with user-generated content are somehow a mainstream product, if only “done right”, and that the trick is how to figure out to “do them right”. To be very honest — and you can check it up on my blog — I used to think like that as well. But the more time passes, the less likely I believe this is going to happen. I can imagine that a lot of improvements will bring more people in. A mobile version, a Web-based version, an iPad version — all those will bring more users in. But will they stay? I don’t think so. Second Life is simply too different from any other online experience, and you cannot “turn Second Life into something that Second Life is not”. Sure, LL can close SL down and start developing mobile apps or embedded apps for Facebook (and become one in a million companies to do the same). They could, if they wish, diversify their product line, as some have hinted that Rod is planning to do. A Blue Mars-inspired appliocation for the iPhone and Android which shows your SL avatar, allows you to change clothes, use some gestures, and chat with other residents — even if that’s all it does — could be interesting, if it’s sold for a dollar or so. That (and similar thingies) might give LL some new income sources.

But it’s not the crucial issue. My point is that LL — and most residents, too — are still emotionally attached to the illusion that Second Life (not the one we have now, but possibly one in the future…) is going to be a mainstream product (like Facebook or Twitter), and that they will once again enjoy the benefits of exponential growth, as soon as LL discovers the right combination of features that enables SL to become a mainstream product.

Well, I think that’s completely the wrong attitude. I’m quite convinced these days that Second Life (and these days there is nothing else that compares with SL; I’m considering OpenSim to be just a variant of SL running under open source software, of course) is simply a niche product. And the longer it takes to Linden Lab to realise this, the worse it will be for us all.

Print PDF