A recent press release concerning ArnoldIT's Google monitoring service piqued my interest. It turns out to be a nicely formatted aggregation page for Google blogs. The most recent five blog entries (or titles therefrom) are grouped together by category. Google has over 70 different blogs (for everything from Gears and Gadgets to OpenSocial and Chrome). Keeping up with them all is nearly impossible. Hence the ArnoldIT aggregation service, dubbed "Overflight."
While handy in its own right, Overflight is not available as an RSS feed. It also doesn't seem to be searchable. So I decided to see if I could mash together my own version of Overflight (tailored to my own research needs), using Yahoo Pipes, the visual Web-app builder.
As it turns out, I was able to cobble together an Overflight workalike in a matter of 90 minutes or so (give or take a bag of microwave popcorn). I didn't have time to aggregate all 70-something Google blogs, so I concentrated just on the twelve developer blogs that are of particular interest to me. My app is on the Pipes site as Google Developer Blogs Super-Feed, which you can subscribe to here.
With my super-feed, you can see the title, description, and content for the most recent 8 blog entries in all twelve Google developer blogs that I chose to aggregate (AJAX Search API, Gears, Gadgets, OpenSocial, Open Source, Mashup Editor, Web Toolkit, App Engine, Google Code, iGoogle, Desktop, and Data API blogs). That's 96 entries total. Actually, it can be less than that if a blog is cross-categorized, since I included logic that removes duplicates.
A tool of this kind is obviously more useful if it allows searching. The keyword-search version is here. (It supports single words or exact phrases.) You'll notice that after you perform a search, a header bar will appear above the results-list containing various links and buttons you can use to subscribe to (and/or syndicate) that particular search. In other words, you can search on "AJAX" and then subscribe to the query as a feed; then you could search on "Google Docs" and subscribe to that query as a feed. And so on.
Is Pipes the ideal way to build Web apps? Not necessarily. The list of things you can't accomplish with Pipes is quite long, and the learning curve (for what you get) is somewhat steep. But it offers a glimpse (arguably) of how some Web apps will be built in the future.
What this exercise really shows, however, is the power of standards like RSS. This is a point worth emphasizing. As Web content becomes more granular, compositional, and personalizable (not to mention more perishable), subscribability becomes a design consideration. Users want to be able to opt into dynamic content. This is a theme I've seen emerge over the past year in the Web CMS world as well as in Enterprise Search, where it's no longer enough just to let users save queries; they now need to be able to subscribe to their queries (or the content generated by them).
Bottom line? Feed-based delivery of content isn't just about aggregation; it's about empowering users -- giving them the power to choose how they want to consume content. That's a subtle distinction that's driving a good deal of change in the content management industry right now, and it's something we continue to watch carefully.