Our work in Second Life is at a crossroads. I know where we have been, I know where we are going; but I am not sure anymore how we are going to get there.

What follows is a set of preliminary reflections based on reading, thinking, practical work in Second Life with Catharina Gröndahl, and a series of long conversations with Lindy McKeown and Steve Bronack.

Introduction

In Autumn 2005 Camie and I started the Marinetta Marinetta Ombro project in Second Life. The project was intended to explore the social aspects of virtuality and to develop techniques for learning that took advantage of those. We created a virtual culture, established Rosario, and explored ways to approach the residents of Second Life as a learning resource.

In many ways the project climaxed in May 2007 with the Semano Semano festival, held to coincide with the Eurovision Song Contest finals in Helsinki, and it was declared finished in December 2008. Since then we have explored ideas of spatiality and geography in Second Life, which included tearing the nine-sim island down and rebuilding it from scratch.

Some of this is documented in Catharina Gröndahl’s thesis Updating Rosario. The rest of it will be documented in a book to be published in 2011. (Note: I will post urls to both of these soon.)

The last twelve months have also served as preliminary research for the programme for 2010-2011. This was conceived as a one year project to produce a simple open-source kit for using Second Life as an alternative to video-conferencing.

We had approached Second Life for an immersionist point of view. Now we wanted to adopt the role of augmentists. We wanted to see if we could produce a simple alternative to Venuegen or Teleplace: an alternative that was free or very cheap to use, and which could be used by small businesses, clubs and families in the Nordic countries.

Because of these plans I have been looking again at various different aspects of the Second Life ecosystem, including Second Life itself, the various official and unofficial alternatives, and the various viewers used to access both. The current position is both more complex and less predictable than I had expected.

Linden Labs: the last three years

Linden Labs was started by Philip Rosendale and launched Second Life in 1999. At the end of 2007 the company went through convulsions that resulted in Cory Ondrejka, the chief technical officer and co-driving force, leaving. He had been the fourth person to join the company, and his departure ushered in a new era, that resulted in Philip Rosendale himself resigning barely three months later

Since then Linden Labs have imported new management and adopted more obviously commercial strategies.

They have downplayed the role of education, which resulted in the disappearance of John Lester (Pathfinder Linden) from the ranks. They have moved the “adult content” into a ghetto continent of its own, and made increasingly strenuous attempts to impose content labelling. They have launched Second Life Grid, aimed at big businesses who want to run their own private virtual world on their own servers. They have attempted to embrace social media by buying a series of web services to create an online marketplace for virtual goods and a social networking site for avatars.

They have developed and launched Viewer2, which was designed to reduce the learning curve for new users by providing a simpler and more intuitive user interface. It was also intended to boost dramatically the ability of content creators to display media in-world by providing an in-world browser, and by making shared media a workable proposition.

Linden Labs: the current state

According to the current CEO, Mark Kingdon, this strategy has broadly been successful, and is continuing to become more successful. Outside commentators have been more sceptical.

When companies introduce an innovative new product, and that product becomes successful, the company almost always shouts about it, as Apple has been shouting recently about the demand for iPads. That is not just because it is good publicity, its because news that something has become popular and trendy is likely to generate an enormous second wave of sales as the mass market follows the early adapters.

From this perspective the fact that Linden Labs have said next to nothing about the success of the world-in-a-box that is Second Life Grid tends to suggest that there has been no success to talk about. Something similar could be said about Viewer2. If it had resulted in millions of new, happy customers then we would not be able to dodge the press coverage.

What press coverage, you ask. Exactly.

To add to all this, on June 9th Linden Labs fired one third of their staff, a move that Mark Kingdon tried to imply was an outgrowth of success. He said

We’ve emerged from a two-year investment period during which, among other things, we’ve spent a considerable amount of time improving reliability and the overall user experience. Today’s announcement about our reorganization will help us make Second Life® even simpler, more enjoyable, relevant and engaging for consumers starting with their first experience.

Interestingly he also said something that I at least had never heard before:

It will also enable us to invest in bringing 3D to the web

This snippet seems to me to be more important than it might seem, because if Linden Labs are really planning to move Second Life to the web, then that raises a lot of interesting questions about where we might best place our limited resources.

Second Life: the alternatives

One of the things that happened during the last turbulent three years was that Linden Labs decided to open source the code for the Second Life viewer and (eventually) the code for the server too. Why they did this is a matter of speculation.

There were two important results of this. The first was that a number of different groups produced viewers with different strengths. This included the Second Life development team who, under Philip Rosendale, created the SnowGlobe project to develop a better open source viewer. There is now a list of officially approved unofficial viewers available on the Second Life site.

The second result was the creation of OpenSim, which is a complete (and almost completely compatible) open source clone of the entire Second Life code base. This originated as a programmers’ hobby but has rapidly developed into something stable and usable by the non-technical.

In part this is because OpenSim has itself developed an ecosystem, and now ReactionGrid have turned up with the aim of becoming the Red Hat of OpenSim. They offer commercial services and hosted worlds that are much cheaper than the Second Life equivalents.

Because OpenSim is open source anyone can download the code and run it on their own server. Lindy knows a number of people who have done this. However, they all say that, at the moment, this is not a trivial undertaking and hosting an OpenSim grid is definitely not as simple as installing WordPress or Drupal on your own server.

So what does ReactionGrid offer that makes them interesting?

The first thing is that, in terms of privacy, their worlds work the other way from from Second Life. In Second Life there is one big world and it is all connected. If you want to build a wall round it then you have problems that you need to sort out. In ReactionGrid every world is separate from every other world, and if you want to link worlds then you do so by a process of federating, using an extra piece of software they refer to as Hypergrid. This enables avatars to move from one separate world to another taking their clothes and inventory with them. The worlds they can move between, however, are limited to the worlds that the Hypergrid is programmed to recognise.

The second thing that ReactionGrid offer is Jibe, which is a newly launched web based virtual worlds system, using the Unity 3D engine. This, they say,

allows us to deliver virtual worlds with the easiest of logins & navigation in browsers, Iphone, Ipad, Ipod Touch, Wii Console, Mac, PC & soon more.

The third thing they offer, for both products, is full import and export facilities. In both you can build models in Blender, 3D Max, Maya, and other software, and save models in the world in formats that these programs will read. This makes the process of building much easier to teach, learn and manage; and it means that historical snapshots of a sim can be saved and backed up in their entirety.

Conclusion

The project we are beginning in August was conceived as a way of utilising Second Life for something different from what we had been doing previously. However it was also conceived as a way of creating a product or set of tools that other people might find useful.

At the time that we proposed the idea it seemed obvious that Second Life would form the backbone of the structure. Now I am not so sure.

Jibe offers a cross-platform approach that makes a lot of sense. Both Jibe and ReactionGrid offer the possibility of creating complex models outside the world itself.

In addition to this Unity 3D is a development platform in its own right, with a basic version that is free. At this momet I have no idea what the advantages of Jibe are over a self-hosted Unity 3D installation.

Erm, guess what I will be doing this summer?