Lately I've been thinking a lot about how to encourage social software systems (including all the soft and hard human processes, systems and technologies that surround the software) to thrive.

Let's imagine, for simplicity's sake, that this site's (the Landing's) inhabitants come in two species: experienced users and new users. The Landing is already populated with a few hundred experienced users, who have migrated from an older Elgg instance to this one. They are, predominantly, students and staff with an interest in online and distance learning. That's the history, that's where we are now on this site.

Now we are trying to encourage diverse uses, bringing in many people with many different interests and needs. I have strong grounds for believing this to be necessary, the simplest and most compelling of which is that the more reasons people have to visit a social site, the more likely they are to do so, making it a richer, more vibrant space with every visit they make. It's a positive feedback generator that works in everyone's favour, leading to greater population density, greater interest, more exciting content, stronger and richer connections. Killing diversity kills a social site.

New users have no connections with others. They are members of no groups and they have posted no content. Unless they come in with a clear and focussed agenda, their primary need is therefore to start making connections, joining groups, making groups, adding content and so on. So, what does the site offer to newly logged-in users? A front page that shows current and recent activity across the site which is, of course, primarily related to online and distance learning. If this is of no interest, they will quickly be put off using the site, especially when they discover that drilling down through different tools actually increases that confusion rather than reducing it. Of course, some will be interested in this kind of thing and will reinforce what is there on the subject. Others will see the value of the system and join in regardless, so we might expect to see an increase in diversity over time, albeit at the very high cost of losing a good many latent contributors along the way.

Meanwhile, experienced users, landing in the same place, are faced with an increasingly overwhelming tide of stuff, much of which is irrelevant to them, which is starting to arise from the tentative first steps of new users who persist despite the apparent bias, or even those who experiment once then leave (a particularly bad case because interaction is not sustainable with such users). Serendipity may sometimes lead them to good things but, on the whole, they are likely to be put off by the large amount of activity. They mostly know why they are here and, while they may have an interest in extending their activities and should be encouraged to do so, an excess of random activity is not helpful.

So - there are two distinct feedback loops at work here: on the one hand, some people are put off by too much diverse activity while others are put off by insufficient diversity. It is possible and desirable that this should reach equilibrium in the end, but it could equally lead either to stagnation through lack of diversity (leading to lack of use), or chaos (leading to lack of use). To make matters worse, the transition from new to experienced use is gradual and there are many in-between users for whom the best a single page can provide is a poor compromise, leading to lack of use.

How could we stop stagnation or chaos occurring? There are lots of more or less top-down interventions we could make, including deliberately pushing certain sorts of content, preferentially exaggerating the novel, adaptively changing the view according to experience, deliberately excluding certain tags from the tag cloud,  and so on. We are indeed doing one or two of those things to help kick start the process. However, If we solve the problem of the front page through active intervention we run the risk of ostracising experienced users, unless they have an alternative jumping off point.

The Elgg environment underneath the Landing offers one way out for experienced users - the Dashboard (a personal and private learning space not shared with anyone else) is highly configurable through widgets and can be used to filter content to a manageable level. Alternatively, people might use their Profile (the public face they present to others) to do the same thing in a more social way. These interfaces both allow the user to assemble the widgets they find most useful and to interact with the site in a more controlled manner, showing more of what they want rather than just what the front page provides. The widgets are mostly dynamically populated and some can show the same kind of diverse content as the front page, but it is largely up to the user whether they are shown and what they show. This would work better if people were allowed to choose one of these spaces as their landing point on the Landing, especially if that choice were to be made obvious and easy, and we might consider making this an option, rather than dumping everyone on the front page. I think there is a better way, however.

In real life:

1. we present different personas or facades to different people, and
2. we have different needs in different circumstances and different occasions - we don't always need a single set of tools, no matter how big that set might be (in fact, too many tools are constraining because it is harder to find them and to choose between similar ones).

These two facts are largely ignored by existing social sites and systems.

For the first case (of offering different facades), a single profile does not help much, even one such as that provided here which allows us to use widgets and permissions to choose what we display to whom. Instead, I should be allowed to make the face I present to my friends to be different from that I show to my students, my teachers, my fellow researchers, the people I work with in teams, the people examining my work and so on. It should be possible for each to look different, feel different, present different content.

For the second case of personal needs, if we are coming to a social site for many reasons, that site should support us in the different activities that each purpose entails. If our interest is currently in working on a course, then we should see a view that assembles the things we need for that course. If we are interested in socialising, then we should see the things that help us do that. Actually, the system shouldn't do that: it should make it easy for us to do that ourselves.

The answer, I think, is to provide a simple tabbed interface that allows us to create and manage spaces for all the uses we need, both for our own interests and to present ourselves to others. For some people I would like to show my research, to others my teaching, to others my social activities, and maybe there are some things I'd like to show to everyone, and I'd like to do that in visually different ways. At a personal working level I'd like a tab to organise my project work, a tab to organise my teaching, a tab to organise my social interactions on the site and, in each one, a different set of resources, connections and tools. As my context shifts, so should the assemblies of tools and sub-spaces available shift. I expect other people would like things organised in different ways and I would certainly want to change the way these things are organised for my changing needs over time. I certainly don't want anyone else to organise this for me, though I'm happy to be given templates for the basics (a whole-site view, a personal view, a public view, for example).

I've concentrated here on individuals, but similar issues affect groups. They too have internal processes and external facades. They too are faced with cold starts and changing needs. A tabbed and controllable interface works for them too.

All of this comes down to a few clear principles:

• diversity is good
• control and personal ownership is good
• different contexts demand different solutions
  
• change is a part of life and our systems should not ignore it
• to support evolving diversity, adaptable parcellation is necessary in any social space beyond the smallest and most trivial

I am having a set of plugins made that try to make use of the tabbed interface I suggest here and hope to try them out soon. Any ideas for enhancements or features would be most welcome.