There are still wrinkles that need ironing (which comes as no surprise to those familiar with my sartorial style), but the Fishtank for Faculty Academy is up and running. Here's a description and screenshot, along with how it works and some lessons learned.

Fishtanks are colorful, interesting, and constantly transforming -- just like the intellectual life around the Faculty Academy. We're using that metaphor to describe what we're capturing here, a view on the blog posts about Faculty Academy events. (Information about registering your blog to be included in the list is in your program). The blog page gets you into scuba gear to dive right into the depths of the blogs. Here, you're looking into the fishtank to see an overview of the posts, including the author, tags and/or categories used for the post, a preview, and what sites the post links to. You can focus in on any of those aspects by clicking on the information in the four boxes below (kinda like focusing on the little treasure chest, or on the log, or on the kelp, etc.). Enjoy the variety of ways to look at our time together (just don't ask who's the plecostomus).

How it works

First, we asked attendees who plan to be blogging the conference to register their blogs, so we have a list of the blogs, who owns them, and their RSS feeds. (The RSS feeds are also used by Jim Groom to use WordPress-O-Matic to collect all the posts to the Faculty Academy blog page). A (somewhat hastily cobbled) PHP script uses RAP and Triplr to look for posts with "FA07" as the tag or category put that info into a big, beautiful RDF Graph. Along the way, I add in some extra SIOC data. This much is reworking earlier work I blogged about here. Then, parts of that graph are exported into a JSON file to be used by SIMILE's Exhibit. And "ta-daa!" you've got the exhibit of people blogging the Faculty Academy.

Lessons Learned (so far)

There's lots here to put in the category of "learning from experience." Here's a few things I'm realizing so far:

To slash or not to slash?

In the blog registration, I asked for the URI of the blog, which then gets associated with the person registering it. Sometimes they've added a trailing slash, sometimes not. But, when the RDF gets collected from the feed, sometimes, too, the channel has a trailing slash, and sometimes not. That's complicated by blogs that exist in subfolders on the server. So sometimes I get a mismatch between the registered URI of the blog and the URI given in the feed, which makes the sioc:has_parent predicates stumble. In future developments, I'll have to use the actual info in the feed more carefully.

Post update times

Because I was most confident in finding RSS2.0 feeds consistently, the script looks for those. Unfortunately, I don't often get info there about an update time for posts, which means that the script is pretty inefecient about what it does and does not aggregate. In future, I should always look first for Atom feeds, which seem more likely to carry that info. This issues is exacerbated by the fact that the hosting provider sets limits on CPU usage that can quickly cause me problems. That's first a matter of my inexperience as a programmer (I wrote my first line of PHP just over a year ago, and it's really in the minority of my time on the job). I might end up doing some quick hacks during the conference to spread out the load on the system.

More lessons are sure to come...