I came across Dapper. This is a fantastic tool that allows you to take structured data on the web and aggregate and present it in the format that you want. So let us take this from the beginning.
We have static ad-hoc web pages that have no real internal structure (like this post content). Content in these pages is a monolithic chunk. However, if I were to give it a structure, maybe I would have an introduction section, a set of objectives that I am going to cover in the post, the content for each separate argument and a summary. If I could make it a structure that could be read by a program (separated into HTML tables with each table / table row having a unique identifier), then I could “read” this structure through a program and regurgitate it out into a different format (maybe a Flash animation or a new way of presentation or a mash-up with Google Maps).
What if I had multiple such structured posts and I wanted to take only the ones relevant to me through a search. I could simply write a program that could filter on my keywords, pick up only those posts that I am interested in, extract out a subset or all of the structured “fields” contained in the post and convert it into a different format.
What if I had multiple sites with multiple structured posts and search features? I could scan all those sites, identify the posts I may want, mix them up together, sort them and display them in a new format.
The good news is that almost all the content you see on the Web 2.0 world is pretty much structured “externally”. That is, for example, a post on a blog can be identified by it’s author, title, publish date and content. You could even get a discussion going (like an email thread, though I have not seen any tool do that yet) on blog comments (and maybe that is why people are gravitating towards wikis or discussion groups to synthesize a discussion).
By “externally”, I mean that this kind of content (a blog post, a shared photo/slideshow/video, a search result etc), is identifiable by it’s outward/external characteristics (who posted it, when, what does it contain etc).
Correspondingly, you could have an “internal” representation as well that dives into a structured representation of the content itself. For example, I talked about the objectives for the post and of having sections of my post connected to the objectives I defined for the post. I could then theoretically find out and aggregate all the objectives (and associated content) from multiple authors and multiple sites, into one common learning store around that objective.
So, for example, I created a lesson called “Dapper” and it has two objectives – Explaining what a structured web page is and explaining how you can write a program that uses structured pages as an input to generate a different set of outputs. Tom, on the other hand, a fictitious author, decided to write a structured lesson, which talked about one of these two objectives besides others. Harry, another fictitious author, had an objective similar to my second objective.
When I would aggregate the three, I could program to search for similar objectives and aggregate all content matching those parts of my lesson. Once I have the aggregated content, I could synthesize and create a better looking set of objectives or learn from what others are doing.
The power is out there with Dapper. Most of the content you see is at least externally structured. This external structure may be augmented by a whole load of other “external” identifiers from other social networking sites (for example, Tom’s post belongs to his website and his website contains his LinkedIn profile link (again structured data) which shows his company and location. I could then mash-up to see what people in Tom’s company are writing and sharing around the world by using a tool like Dapper to create the program to perform the actual linkages and presentation of the synthesized information.
Another tool that merits a deep look is Sprout. In it’s beta, it offers a similar capability to syndicate content and present it in different formats.
This capability would, in my opinion, be a fantastic addition to any PLE interested in being able to allow learners to explore, locate and create views around their learning information sources. It is a semantic web, isn’t it?