Content Pipelines


Clemens Vasters is talking about content pipelines. Clemens envisions a system for processing RSS feeds, Exchange folders, public websites and other information that pushes this content through a series of stages in a pipeline that processes the information in specific ways. For example, one stage might add Google search results to the content and another might annotate it with relevant books from Amazon. Still others might perform language translation or dictionary links. Clemens sees this as an open architecture that allows stages in the pipeline to be written by anyone and routed to as necessary. In other words, this is a set of web services for RSS feeds. This is a cool idea that generalizes the Mockerybird Book Watch. Clemens says:

So,╩if such a distributed infrastructure existed, and you'd aggregate this entry "backrouted" through a pipeline of filters provided by Weather.com, Google.com, Dictionary.com and Amazon.com, you'd have the weather for Athens and Madrid, all relevant Google links and books on╩"content" and/or╩"pipelines" and WS-Routing, and╩links to╩explanations of all non-trivial words in this text. How's that?

The lesson here is that its the links to related information that are adding value. For example, for most of the words in a typical post, I don't need an artificial link to the definition, because my brain has already made that link internally. For some words though, it would be helpful. One of the interesting aspects of this project would be the user interface issues to ensure all this information make sense once it gets to the user.