SXSW: Chris Messina and Actvity Streams

LIVEBLOGGED – so pretty much as said

Google Data Liberation – most excited to be part of this team

Ingeneral, interested in generative systems and structures, rhizomatic structures, built into the fabric of how they work. Start spill with constructs that grow into the systems. Thats how it all started, hashtags etc.

As in The Future of the Internet, (Zitrain) those sort of systems, paints a picture of things we need to think about as internists. So all that I have been doing based on these generative systems, those with transparent DNA.

An activity stream – facebook newsfeed, we are going somewhere with this, a lot more interesting. In 1999, Fight Club presented us something that gives us something to think about. YOu get a picture of someone through material objects. so imagine going shopping you get an activity stream as a receipt of what you have purchased. YOu get this on paper now, so how do you get it to something that is usable. When you purchase things with cards, the bank does all the collection and analysis. Card becomes a digital identity and you have no idea what the digital life is like. We end up with a stack of receipts that you can’t get anything from. Ther is an unbalance, where those that provide the cards can get the info and make use of it and they do not give us anything back. The bank provides data (he asked for bank info) but they provided it as pdf data. The pdf is a digital version of the paper they used to send you.

Today, the newsfeed is the best activity stream there is, one of the only ones we can have. (History of Feeds) IN 1999, they took RSS and piped data form one place to another. RSS was title + link + description. this was when people were afraid of giving data away, so this was the most you could do. So go forward 5-6 years, then there was a series of battles, so in 2005, there was a new format, called Atom. Still similar, till about syndication, innovation was author, unique id and update date. So very slow to get there. now we know who wrote it and when it was published. Moving forward but not about powering social web, still based on media consumption, the syndication. So now, we take the articles, will still have the same idea. We have a news feeds, which is still a portal types thing. So even with the most advanced browser/feed you still get the news feed. We are trying to pump all these rich info throughout he formats that were designed to publish news articles. So everything looks the same, it is all RSS (etc). There are all these rich activities but the formats we have are all stuck in 1999. The social web needs to have better, richer formats to allow us to express, why, when, how.

This is the friendfeed problem. There idea was to bring them all together, to make sense of it all. It was a metalled, representative of your actual identity online. Then friendfeed got acquired by facebook, now the service has languished. it still supports services that do not exist, it does not list all the services. (So no work). Int he world of social web and startups, there are a lot of casualties. Friendfeed put a lot of effort into stuff that goes away (as would anyone). So the solution could be a universal format (to minimise effort).

So this is where activity streams format comes from.

So to start, let’s go back to the Soviets. They proposed a theory called Activity theory…a structure about making workers more productive, a system to create divisions of labour. You wool have a subject with Tools with an object to produce and outcome. Then the theory got expanded, so broaden it out to the individual operating in a community structure, with rules and roles and mediating artefacts. This allows us to think about activity streams beyond point a to be syndication. It allows us to create meaning. (Engestrom 1967)

one of things you can think of is Social Objects. Jyri took his Dad’s research and started working on social objects. You can’tt just do a social network that allows friending, you have to create these shared objects that people can gather around and have some interesting interactions. Those social objects are the pivot points for connections, which allows them to derive meaning and make sense of the connection.

You Tube – the social object is the video. Instead of a list, you can modify it, with comments, favourites etc, all this social residue provides meaning, adds value to the object, that was created once and then gets built up over time. Look at Flickr, go back to our activity theory, they have understood how the different things they have used these rules to generate interesting interaction models. Focus on the pieces, you create a vibrant community, different roles, create, comment, curation,. Flickr does a good job of expanding the roles of non-producers, the actions adds richness and dynamism to Flickr than found on other sites. Also, on rules, they have made it possible for an individual to finely tune the system, the rules. So if I were private, I can control public searches. When I get a comment from someone I don’t know, then there would be a connection as that is the only way to find. If I was public , then another set of reason for people to find them. So the way I set up attracts the activities that are meaningful.

Useful when talking abut lifestreams. Lifestreams and activity streams are not completely interchangeable. The lifestream concept came from David Gelernter. Wrote Mirror Worlds. A decade later 2000, wrote The Second COming, A Manifesto. He talked about lifestreams. He said a lifestream organises info as a mind, not a file system. He talked about the idea of hashtags, connecting elements. (HTtp://j.mp/gelernter) This is all like Donnie Darko – we have these threads that play out to past and future and we can modify, but we do not have great tools for doing so. As we start to move towards experiencing the now, the next and the next next.

As we start to produce all these digital objects, we start to snack on it. Today may feel like overload, we don’t have the tools to consume it. We are constantly compressing, microcontent goodness. You would write a book, long, you were paid on words. Then we went to articles in publication, now we do tweets. So what are we giving up? We go from a slow consume and digest, lingering on content to one where info is disposable as there is so much of it. What is interesting is the data trail that can come from these experiences.
The info can lead to a mapping of behaviour. We are scratching the surface of this, with systems about where and what we are doing in the world. Still not good systems. Being able to build a profile over time, this social data, social residue and be able to make sense of it is fascinating. We can make little of it now.

Social signifiers can be useful in training computers to serve us better; it is very valuable, and bite size chunks make it more accessible to computers. One example is a pedometer I use, that hooks to a website, so I can track the trend. You need to accrue data over time. So I can track have a slowed down, am I more lazy than my peers. Look at the Feltron annual reports. He collects all of the data over a year. and publishes. He tracks where he goes, so why can’t google maps provide this? Food consumption, how does he track it? There’s an app called Last History…looks over the Last.fm scrobbling. This is your data, so you can do something with it. It takes your habits, combines ical, iphoto, imovie etc and shows you the soundtrack of your life.

Now Tufte has been horrid by Obama, to track and visualise where our tax dollars are going. So why don’t they make it compulsory for the data to be released so we can build stuff….we’ll end up with pdfs from most of them.

The solution to data overload is more data – actually more metadata, data about data. So we need to start generating this and this is what we are doing with activity streams. We presume there is an actor, that did something and modified an object, with some output/target. Actor Verb Object Target. this model allows me to do this.

So what does the code look like? So add verb, object-type and target to the atom definition. so you have an activity stream data model. The new bits are added. You can start to substitute verbs and object types, mix and match and build more interesting experiences. Now we have a list of verbs and object types, from review of friendfeed and others. So all of this is in v0.8, moving towards 1. the idea is not to have a million verbs on day 2, but to start with something that can be added to.

We are not really inventing this, we are being inspired by the microformats process, for the expansion of the model. So ask why, do your homework, then propose the new verb/targets etc. Then iterate. Then interoperate. We want to grow this slowly so we know what we are getting into

http://activitystrea.ms