Information Cartography

Tangent - Managing large content data sets and individual state

An interesting discussion recently provoked thoughts around an earlier idea, presented below.  Other than using content replication system (CRS), how does one manage state with respect to large web sites?

As most content publishing systems, editors or user determine a document (content) state.  State is described as paradigms that follow: idea, draft, review, published, retired.  Current systems use workflows that are user initiated to determine state. 

Document state should be determined by the document itself as non-cooperative zero sum game theory.  Non cooperative game theory has been applied to the field of economics.  However this can be applied to documents as well. 

As an user generates an idea for content, that idea is transformed into a draft document.  That user writes a draft.  Then another user either accepts that document to be published or send its back to the user for more revision.  Once that document is published it might retire or be moved to a different location.

This leads to the following generalized problem:  1% of the content is viewed/consumed by customers 90% of the time.

Link to the lab site

The first prototype is out - but read this first!

Does not work in FireFox, Safari

To use, simply just double click on System to zoom in or drill down, just are you would when using http://local.live.com.  Mouse scrolls work also to zoom in and zoom out.  You can hold the mouse down to scroll in all directions.  The higher the zoom level, when using a mouse scroll wheel, causes the API to go the left of the screen. 

So, have fun with this.  Based on the feedback we will try to enable other Virtual Earth features in the next few months (creating routes/relationship between apis).  Think about what you see and give us feedback. 

Here is the link, http://labs.msdn.microsoft.com/vam/

Introducing..

How many times do you wish navigating the Table of Contents on http://msdn2.microsoft.com/library/ was easier?  How should we expose the visualization of navigation a tree?  Is this even the right layout?  How does the user find common searches that might solve their problem?  How do we expose code snippets? 

What does the world look like when there a collision of Microsoft Virtual Earth and Microsoft .NET APIs?  There is ability to override tiles in Microsoft Virtual Earth, allowing any type of information visualization to occur.  Imagine this - you go to a web site with an image of your car - you can click into the car, zoom in/out, pan around, etc.  The first layer is the image of the entire vehicle.  As you zoom in, you see the parts of the vehicle.  Add mash-up data, and now you've got car parts...  Sure - there are lots of different controls out there that do this - but thats the point - why download another control to the user?  And of course, someone has to know how to use the control, add the data, etc.. etc..  Why isn't it easy?

So, instead of the automobile example - we are rendering APIs instead.  Here is a screenshot from a prototype,

The problem around navigating the table contents might not be solved but here is a different approach, which will be coming out soon on http://labs.msdn.microsoft.com/.  This information was presented at CSCW 2006, at the Mash-ups Workshop.  Here is a link to that presentation.

Help us build a better experience. In the upcoming weeks, we release a limited version of the application on our lab site. Send us comments through our blog.