SXSW: How to Leverage Solipsism

Veen : intro
Agreed to do panel because been working on tools in blog world.
Worked with Technoratti, Blogger, Creative Commons.
Seen trends coming up.
Set foundation for discussion.
What does solipsism mean?
"Nobody exists but me!"
Understanding of self as only thing in the universe.
What does that have to do with IA?
If I think about my info (my documents folder), it's a self-centered org.
It doesn't matter if you can figure out my folder, only matters I can. - can surf and collect bookmarks.
You can tag them, which is like making folders in My Documents.
I can click on blogs and get my sites about blogs.
Related tags shows intersections of tags.
At bottom of list: "blog" from all users.
Can view intersection of my private info with the public domain.
Can use this to discover people with similar interest with my links metadata.
We have had social networks for a couple of years.
Everybody was talking about Friendster a year ago.
Social networks are kind of like a class yearbook: you pass it around, get it signed, then put it on a shelf and never go back.
Combine friends network with photos you take, you get something interesting like flickr.
Posted photo of lizard in San Francisco park, can tag with keywords.
Trends start to emerge from keyword tags.
We can see what people are tagging in a large sense (300K people on flickr).
Then add an open API, you can do things like apply geographic coordinates.
Is it useful? Not sure. Is it interesting? Very!
Big sites do this too.
Amazon -- not as interesting but maybe more valuable.
It watches your behavior and customizes the site.
This is the "candy at the checkout" impulse buy at the supermarket.
The thing that overarches all of this: may be interesting on a case-by-case basis, but what happens if you find yourself with accounts at who-knows-how-many places.
What happens with all that personal information in all these places?
How to you manage it? How do you control it? How do you get it back?
Had to send four emails to get info out of linkdin.
Can Travelocity know my schedule so I don't book at wrong time?
Can I import friendster into flickr? What are the implications?


Vander Wall : how many have emailed info yourself? (lots)
How many have copied or pasted from web page or word to doc to use later? (lots)
What we are looking at his the personal view.
The info that comes in is our view.
Our view of the world is how we organize the room on our side of the window.
We have our own organizational system behind the window.
As people come to our window, they see our organizational system.
What we are trying to do on the web is we have a view of the information, but we have poor structures to organize on our side.
The world of paper is falling over and being replaced with digital properties.
Digital rights management: written word was always meant to be used, DRM may make it unusable.
Our world is made up of digital information, and information shapes us.
Our digital personae is different from what we are in person.
Shapes how other people see us, and how we see the world around us.
One problem with information is we get lost early.
There is a huge morass. Went from scent of info to morass of info.
We can search and find, how do we re-find?
Google has changed, Amazon has changed category structure.
Blame me for the term "folksonomy".
Usability is king, but refindability is the prince.
We have our own personal language on how we get back to things.
We have silos of info around us, how do we jump between disciplines?
Do I look for "human resources" or "employee benefits" to find the vacation schedule?
I want to find my information in my own vocabulary.
That "syncing" feeling.
We have devices with information, and we have problems getting things to sync.
Information is lost or overwritten.
Different things on work computer, home computer, laptop, Ipod, pda, etc.
Need some mothership to hold all the information.
Need versioning for the information.
An RSS feed for every document I have?
How do we make syncing easy?
How do we make information externally controllable? What are the trust networks?
Building a personal information cloud.
Key attributes: portability/ubiquity, access, personally organized.
Portable - can I access it while I'm on the beach.
Access - will my device be able to access my information.
Personally organized - it is how I think. It is how I call things.
External storage and management are important.
Flickr,, other personal portals manage our info.
Screw that, I want one portal that I control to manage my info.
Share it out to everybody else.
attention.xml - A Technoratti project. Manage massive range of RSS feeds.
Also manage the info on my hard drive.
Personal storage - able to have access on your hip.
Standards - open APIs are insanely important.

Turnbull: "I'm from the university and I'm here to help."
Things are hard to classify.
A lot of theory people may think flickr is anarchy.
"The inmates are tagging the asylum."
Academics may not like it, but are starting to use this style: CiteULike.
Use of tagging: collaborative filtering, recommender systems, information filtering, search system augmentation.
Let's make this focus on the user's view, not how a system should represent this.
There are some interests in this.
How do you get people to cooperate? How do you get them to share?
How good can tags be?
Systems are failing with the scale we have now (Netflix says "no recommendations available").
Buy a kids book from Amazon as a present, you keep getting recommendations for more even though you don't want any.
The web is a shared info space, but there isn't much sharing. The potential for this is large.
Tag issues - spamming and gaming. Will there be a tag optimization business?
Tags are explicit.
Tags are text and so can be analyzed. Extraction. Synonmys.
Sorting out tag properties: power law distributions.
Supertaggers - people who are prolific taggers.
Systems to find "our critic", supertaggers that meet our sensibilities.
Social issues - who controls the sharing?
Do you promote with "give to get" systems?
Anonymity versus community. "People are great data points."
Issues with free riders, people who never tags.
Community centered tagging.
Useful metadata. Implicit tagging - tags based on what you do.
Amazon remembers magazines you look at, offers trial subscription.
Don't Fence Me In - I want tag mobility. Don't want to be locked in.
If I want to go from Netflix to Amazon, how can I move my data? (You can't.)


Celik: Amazon keeps track of what you buy.
Was looking at recommendations. Can rate things and say "stop recommending."
If you buy off somebody's wish list, Amazon things you own it, even though Amazon knows you don't.
Getting back to tags.
Technoratti tags.
A "tagsplosion" happening. Multitude of services allowing tags.
Wanted to do something, not sure what.
Technoratti isn't really a store of data.
What can we do that isn't storing.
"New Year's Resolutions 2005" - bloggers link to a page and it shows up.
People started just linking to the page saying "check out cool site" without resolutions.
Linking as a tag versus linking as a reference.
Introduced rel="tag" to mark a link that is a tag.
Technoratti indexed metadata such as DC.subject.
To make it easy, added link to tag page.
Or link to any tag space.
Big difference between Technoratti and what other sites are doing.
Technoratti tags are on blog posts, doesn't hold them locally.
If Technoratti goes down, tags are still there.
Tags are visible.
What happens if you take these tags spaces and mash them together? Do you get interesting results?
How can everybody keep their own on their blog.


Veen: I used to design web interfaces to be flexible and lightweight.
Architect to be able to turn on a dime.
These new systems are about no architecture.

Turnbull: people have been doing classification research for 86.2 years.
They are going to be gazing at navel for another 86.2 years.
Let them do that, I'd like to go ahead and build things.
It's not perfect but better than nothing, and we have a whole lot of nothing now.
Info is a little more complicated than the Yahoo taxonomy.

Veen: people were saying "how is html going to work, it doesn't have a central server"
It feels like we are going through that now.
What do you see in research how people are categorizing personally?

Vander Wall: When people realize tagging is social, not just for themselves, their list of tags expand.
Will go from 20-30 when tagging to themselves to 120-130 when they discover friends' tags.

Veen: can stuff be too easy? Tags or metadata become too broad and worthless.

Celik: we are far away from that being a problem.

Turnbull: could over-specialize.


Q: using tags in citizen journalism footage.
Did filming for documentary.
Wants to form an on-line social network and allow people to tag body of interviews.
How do you associate meta-data with tags? Such as identity.
"I want to eliminate this viewpoint."
"I want to tailor to this audience."

Veen: Restate: How do you assign validity to tags? Involve identity.

Celik: anything you can blog you can tag.
If you want to keep it all on one site, it's up to you.
The unreliable system (web) had better growth characteristics than the reliable ones (gopher).

Q: When is my data to share? When is it private?

Celik: Services that allow you to own and control the data will offer a better experience.

Vander Wall: We need to get more granular with our trust.


Q: Open space technology (meeting technology)

Veen: that's why I have a blog.
I can make information available, start a discussion.

Celik: people are doing conferences saying "the tag for this conference is X"
People blog using this tag.
Allows people to converge easily.


Q: Talk about ownership of data and active tagging. What about behavioral data? Do you own it?

Veen: Difference between I tag and share and I tag and someone's watching?
Those are very different things.

Vander Wall: Came up at Montreal IA summit.
A panel on "why amazon is wrong."
We should own that information: wish list, purchasing info, etc.
Should be able to go to Barnes and Noble and get recommendations.

Veen: open APIs are a trend that are happening, but not fast


Q: What about privacy and security? If we have aggregation across everything, how safe is relationships with businesses, etc.

Turnbull: Even if tags make search smarter, I don't know if I want to share it.
There is no groundswell to ask for this kind off anonymity.

Technorati tag:


Comments have been closed for this entry.

re: SXSW: How to Leverage Solipsism

Thanks for the notes.

re: SXSW: How to Leverage Solipsism

Great notes! For anybody wanting more, here are...

More notes by Liz Lawley

More notes by Josh Greenberg, including links to presentation slides by all panelists