Aggregating a large class

I started out the semester planning on using my own blogging server for my 360-person class. The advantage was simple: I had set it up so that when students set up a blog, it was automatically included on a lilina-based aggregator. I had used the lilina aggregator in earlier, smaller classes, and liked its combination of “river of news” presentation (in which entries are listed in chronological order, no matter what their source), and the ability to hide or reveal entry text using a javascript link on the page. Great stuff.

Then I realized that my poor ISP would drown under 400 new blogs, and decided to outsource. I recommended my students set up on, though some have opted for Livejournal or Blogger. I then set up a page that would allow people to add themselves to the lilina feeds.

This seems easy, but given most of the students had never blogged before, understanding what an aggregator is, or what the URL for their blog is, let alone the URL for their feed, was asking way too much. Even more than a month in, I have a feeling many students have a shaky idea of what these things are.

Fairly early on, it became clear that lilina wasn’t going to be able to handle the load. The problem was that it was set up with a kind of cheap cron: it cached the page for an hour, but if you were the unlucky visitor who showed up after that hour, you would have to wait twenty minutes or more while it went out and checked feeds. What a mess. As a result, it wasn’t being updated, or was timing out when trying to get feeds, and students were panicking because they didn’t know if their blogs could be read. Yes, I could have scheduled an update in the normal way, but my (insert not-so-nice word here) ISP for the site only provided cron jobs that could be executed once every 24-hours. Bad news.

So, I figured, no problem: we’ll move to Bloglines, my favored reader. Nice idea, but with 400 feeds, it remains slow. Moreover, students were now used to the river-of-news style (rather than the folder style), and had trouble figuring out that Bloglines could offer that, too. Now we were two strikes down.

I spent a mad few days trying out every possible combination of web-based aggregators and a number of server aggregators. Maybe I could use a service that would blend all the feeds into one, something like Feedblendr. Nice idea, but didn’t update frequently enough. I worked my way through the list of aggregators on Wikipedia, feeling a bit like the three bears. This one had an intuitive interface, but no way of making it public. Another one was set up for public aggregating, but presented the feeds as folders rather than as a River of News. BlogDigger looked just about perfect, but didn’t seem to actually work.

In the end, I installed Gregarious on a new server (where I could do a 20 minute cron job to keep everything up to date). It’s been working well. There is the small issue of the read/unread items, which doesn’t translate well to a public aggregator, but otherwise it is quick, and fairly intuitive.

Some of the postings so far this semester are outstanding, some are pretty atrocious. With this big a class, my ability to improve the writing of the majority is pretty limited. My push over the next few weeks is to highlight some of the more common grammatical errors, in hopes that we can at least limit these a bit more. I’m also going to put up a tutorial on how to get started on Google Reader, so that students can import the OPML from our aggregator and personalize it on their own private reader.

This entry was posted in Uncategorized and tagged . Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.


  1. Posted 2/23/2006 at 4:02 pm | Permalink

    Alex, this came out today and it seems to have the speed you need… it supports column view, cloud view and river of news style. Dump an OPML and let me know how it works for you…

  2. Posted 2/23/2006 at 9:03 pm | Permalink

    Hmmmmm, sitewide feeds for WMPU sites, run through feed2js = problem solved.

    A la:

    Now, all we need is for WPMU top have an option to ‘represent’ a blog on another platform.

    V.interesting post BTW, confirmed many of the things that I’ve been thinking about of late, thanks for sharing.

  3. Posted 2/23/2006 at 11:22 pm | Permalink

    James: But you’re aggregating the whole site, yes? I think there is a need for aggregation at the class level.

    I don’t know if I detailed this in my blog, but I took a run at WMPU and wasn’t happy with it, I think in part because it limited the flexibility of students who wanted to tweak their install in various ways. Of course, that’s the price of security, and my site has serious security holes. But then, I’m leaving the provision-end to you, since you seem to have a solid handle on things.

    Now that I am *not* focusing on maintaining a centralized site, I do see the advantages of the distributed approach. Time-permitting, I’ll write a script that will allow for multi-gregariousness, as well as secure ways to add blogs (both by an admin and by students). I agree this isn’t as clean as the automatic aggregation–and that may be a good way to go.

    Actually, there is a third route. Using, or whatever, a student could sign up for a blog at the aggregator, and the aggregator could post that information on to, etc. I wonder if that automatic account generation–even if it is tied to a particular email address–would violate their ToS. I have a feeling that they might be OK with it.

  4. Posted 2/26/2006 at 9:49 am | Permalink

    Wow. And thanks. I’m about to set up 120 students blogging with the assumption that they would continue with their blogs for their entire time in our school; i.e. ramping up to 500 students. We’re trying blojsom because it is supposed to fit in with our university’s authentication system, but we’ve not gotten far enough along to consider syndication. Eeep.

One Trackback

  1. By Alex Halavais » on 5/2/2006 at 1:32 pm

    […] At the beginning of the semester I wrote a little about the technical issues of aggregating content for the nearly-400 person class I’m teaching. I finally solved the problem (mostly) by using a copy of Gregarious. (That aggregator is set up here.) I’ll write shortly about how I think that the big blogging course worked out pedagogically, but first a quick technical note. […]

Post a Comment

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>