We are happy to announce our latest launch here at Strands. We are opening the private beta for Strands.com, a very early version of a destination site that aims at helping people discover new stuff. Our vision for Strands.com includes two innovations towards which we have been working on for quite some time:

1.- Applying our social recommendation technologies on top of the lifestreams of users and their communities, to provide them with personalized recommendations of media content they might like, based on their behavior.

2.- Allowing data portability for taste information, enabling users to take their taste profiles with them to other sites.

When we are online, we generate at least two types of data. Digital assets such as list of friends, pictures, etc, that are normally the focus of the data portability community. And taste-related data, data which describes our preferences, our tastes. This is data that can be encapsulated and shared just like digital assets, and is the area where Strands is focused.

Without inventing new or proprietary technologies, we will identify and encapsulate taste-related data and enable the users who own it to take it with them. We will leverage into existing efforts including OpenID, OAuth, and APML-style representations for our relative preferences of all types and not just our attention preferences. Users will be able to take their taste profiles with them to other websites that are prepared to be taste profile consumers.

Strands.com will lead the way by aggregating users taste data from the sites and services they choose, and creating a portable taste profile out of it that users control and own.

We expect to roll out these two innovations in the near future, and really look forward to working with all of you to enable open taste sharing across the web. If you’re a website developer and you want to be part of the open taste sharing experience, please let us know.

As a first step towards this goal, we are opening up the Strands.com private today which currently includes:

1.- Filters and Hot Posts - Strands.com helps people cut through the noise of lifestream posts by providing filters, including date, content type, and groups created by the user. For example, you can find what videos your coworkers have favorited this week, or see only what music your partner has listened to today.

Personalized ‘Hot Posts’ highlight what’s popular with your friends, to help you discover new things that your friends have enjoyed.

2.- Granularity and Privacy Options - Strands.com allows you to control what you share with whom. For example, you can post a photo you liked, allowing only your closest friends to see it, while sharing your taste in music with all of your coworkers.

You also have granular control over what you follow from whom. For example, you can follow your friend but filter out his tweets.

3.- Strands Friend Tracker - this desktop application shows you real-time updates on the activities of your friends right on your desktop. It also records your activities – such as listening to songs in iTunes - and posts them to your lifestream.

We will be adding improvements and new features to Strands.com quickly as we move toward our vision of implementing personalized recommendations and open taste sharing.

There is an excellent post by Marshall Kirkpatrick at RWW, discussing the economics of data portability.

“We argue that it’s in everyone’s best interest that the data be freed. Vendors have far more to gain by working to add value to freely flowing data than they do from trying to horde as much data as they can”.

We at Strands have already explained our vision of why we think not only users but also vendors will benefit from data portability, and why it is important for web personalization.

When we are online, we generate at least two types of data. Digital assets such as list of friends, pictures, etc, that are normally the focus of the data portability community. And taste-related data, data which describes our preferences, our tastes. This is data that can be encapsulated and shared just like digital assets, and is the area where Strands is focused.

(By Dr. Rick Hangartner, MyStrands Chief Scientist. Guest column published in ReadWriteWeb)

Fifteen years or so into the evolution of the web, we already have many of the key ideas and technologies in place to start describing and sharing personal preference information — or what we might colloquially call “taste” — in order to personalize web experiences.

So, why haven’t we seen widespread adoption of web personalization? Much of the answer seems to lie in the fact that user expectations and online business models haven’t yet evolved to the point that user-controlled, ‘open taste’ sharing is a viable option.

However, the dataportability.org initiative suggests we may have reached a turning point. The DataPortability project taps into the strong conviction engendered by the do-it-yourself nature of the web 2.0 movement that individuals should “have control over their data by determining how they can use it and who can use it”. This extends to an inherent belief that that it has not been a lack of effective technology that has held back this new culture of open data sharing, but rather, business models that have been over-reliant on laying a proprietary claim to some portion of that data.

Taste sharing is a DataPortability use case
We express our online tastes anytime we make a choice between the various alternatives available to us. Some of our choices may be characterized by the number of times we select each option, for example the number of songs of each genre we play when we select music, when repeatedly confronted with the same choice. Other choices may be expressed subjectively, such as by assigning one to five stars to movies we watch when we are asked to rate our preferences for the different alternatives. In yet other cases, we may in effect be giving estimates for the number of times we would expect to select each alternative, such as when we are asked if we are likely to buy a product or vote for something. Virtually any online experience we have includes one or more instances in which we make conscious choices reflecting our preferences.

For the more theoretically inclined amongst us, we can view a choice as somewhat analogous to a random experiment and our relative preferences as measures of the different possible outcomes of the experiment. The collection of such experiments that we participate in as a matter of course in our web experiences paints a vivid picture of our taste. For the more pragmatic, each time we make choices, we generate data which empirically describes our preferences. This is data that can be encapsulated and shared just like any other picture, blog post, video, or other piece of online content that we create, and which the DataPortability is focused on.

Just a few ideas for open taste sharing
As a DataPortability use case, open taste sharing embodies and embraces the culture shift that the Web 2.0 movement represents. With regard to data ownership, the DataPortability concept has even more succinct expression: our tastes should be ours to share - or not. This puts the user in control of their online experience, so they can set the boundaries of how much they want to share, and with whom. Similarly, there is no need to invent new or proprietary technologies to simply identify, encapsulate, and share taste-related data. A little thought by websites about how to identify and summarize our relative preferences on their site, along with OpenID, OAuth, and a little task-specific XML for markup is enough to do the job. Clearly this kind of data sharing also raises new privacy concerns that will be part of the work-in-progress for the entire DataPortability project.

Perhaps the most interesting challenge lies in educating businesses to thoroughly and thoughtfully examine their current ideas about user data, so they can better understand and enthusiastically embrace The Web 2.0 Golden Rule: “Do for other web experiences providers as they would do for you — under our control as the owners of our taste data — and the blessings of networks effects for taste data shall be yours.

Two of the key trends for 2008, Open Data and Personalization, are rarely considered together, but we have always thought they were strongly related. So when we were asked to talk at Startupalooza (a really cool Portland tech event, put together by Todd Kenefsky and the Legion of Tech) we decided this should be the topic of our talk: taste sharing for web personalization… something which is of extreme importance for MyStrands and the entire recommender industry.

We will write about it more over the next few days but we wanted to share the slides with you first.

| View | Upload your own

Rick Hangartner, Chief Scientist, MyStrands