Posted by Suw Charman-Anderson

[This is liable to being the only talk I take notes on - just too braindead to do more.]

The word ‘beta’ used to mean something. Before Friendster, it used to mean that something was in testing, but now it probably means ‘not yet profitable’.

Software dev used to be a hideous process with specs being written, and checked, and coding and lawyers. But developers don’t like this sort of process.

MySpace developers decided to hack something together using Cold Fusion. No spec, no qa, no usability, no legal, no marketing. They just deployed it. From idea to deployment was two months. Can say ‘maybe they got lucky’, but that’s not the full story. But when they shipped, they asked the users for feedback, and built out requested features. But still no designed system, it’s just piecemeal hacks.

The beta is still pretty standard way of doing something. But MySpace still doesn’t have QA, instead they launch 2 or 3 new features a day, and hack on the live servers. An extreme idea but it’s a new way of working that’s pretty consistent to the social software world.

- hack it up, get it out there
- learn from your users, evolve the system with them
- make your presence known, invite feedback
- monetisation? Add a few ads here and there

Pros and Cons to this.

Cons - produces terribly horrible code that fails frequently. Held together by voodoo. But if you think about how usability been done, has a mentality, lab-driven context. They show people software, ask them to interact with it, and the result is that they fix a few pixels, change working, change page flow. Great for human-computer interaction but not human-human interaction.

But when you have crowds it is different to when you have individuals. Anything that can be fucked with gets fucked with. No good way of testing it per se, but no way to say ‘how to make certain that this can be used. But this is key to what makes social software valuable - it ends up reflecting the crowds.

Because social software spreads friends to friends, this is great from a marketing perspective. Shapes way people use it, the way they think about it. Early days of flicker, Caterina and Stewart said hello to every new users and talked to them about the site and why they did it. That shaped the Flickr culture, and the community because then friends of Caterina and Stewart did the same.

Lots of social software is tech-centred technology.

But an example that went awry. Orkut. Known now as a Brazilian site. WAs originally deployed as an invite-only to tech people, engineers, people associated to Google. But people joined because they knew those people, but weren’t necessarily interested. Too many YSANs.

But then some Brazilians joined, and the list of countries had a list of flags… like a sports event where you want to rise in the rankings. The Brazilians thought this was fascinating because they could beat the Americans. Messages went out and the Brazilians joined en masse, went from 5% to 90% of the people. At this point, the developers weren’t living in the system anymore.

Orkut has now spread to India, and that was deliberate. No problems launching but has taken on a new form recently. The space duplicates the caste system in some detail, and again Orkut does not know what’s going on or how to deal with it.

Culture’s provide meaningful context that tells us how to act. Spaces set norms, e.g. getting on a bus. As kids we don’t abide by norms, until our parents teach us. We learn from people around us and the space itself. Know that a bus is different to an opera house.

So how do you make meaning of context on a social site? Early Usenet groups all looked the same - so how do you make context? Can’t tell the difference between the others except by interaction.

So look at Friendster:
- Gay men
- Bloggers
- Burning Man attendees

Depending on which group people joined Friendster in, they took on that role. So if you were invited by a Burning Man attendee you thought it was a Burning Man site, so you they did their profile as per their own context, and didn’t realise or assume there would be any other contexts but the ones they knew.

Didn’t take long before your boss got online and you realised this was not a place to be half-naked.

Academics often talk about context and what’s appropriate. But it’s very difficult to cross contexts. People change, for example, the way that they speak depending on who they are talking too, but you can’t speak to multiple groups simultaneously without having to make a choice.

When researchers created ARPANET they were interested in sharing information. Interest driven starting point. Even MUDs and MOOs were activity driven.

But inversion with social software, because they started people-first, because you don’t want to know everyone on the web. You may be blogging to reach an audience, but you don’t have billions of readers, you start off with your friends, and if you’re really popular you’ll get beyond that. But your friends build the context. The people you know, connections of relations you have in the system.

Radically different process, but we don’t know how to deal with it. Problem is scale. These contexts collapse. How can you deal with multiple contexts because you’ve scaled.

Monetisation is forcing a lot of sites to scale too fast.

Facebook, for example. Colleges… then… schools… then businesses… now everyone. Tension with marketing - marketing means scale, and scale means multiple contexts.

Delicious, conversations about how awful it was that non-techies were posting.

How far can things scale? Can it work?

Blogs are the only area where its scaled successfully. Because there is no ‘blogosphere’ because it’s not one thing. People are unaware of each other - foodies and knitting bloggers don’t mingle.

Three questions;

Designers: There are costs to chaotic processes behind the design now, and what are the processes that can support users without burning out designers
Researchers: what implications does all this have for society, design, every day life, globalisation
Business folks: Monetisation and growth are seen as desirable, but they destabilise most social software and kill communities. is it possibly to monetise without doing that?

blogtalk reloaded, danah boyd