I’ve been monkeying with a Drupal site that looks like it could fulfill most (even all?) of the mythical Eduglu concept – a website that aggregates all feeds published by students in a class/department/institution, and helps contextualize them in the various groups/cohorts/courses each student participates in. It’s getting really close – it can currently suck in all kinds of feeds, auto-tagging items, and even lets students create their own groups and associate feeds with them. There are issues, to be sure, mostly with respect to honouring the original tags in the aggregated items, and with taking advantage of the social rating system added to the website, but it’s so close I can taste it.
At the moment, there are almost 1200 items aggregated from feeds published by 19 users. It’s only been running for a week, so that’s not a bad start…
One added bonus of using Drupal for this, is that I can drop the Tagadelic module into place to generate a tag cloud representing all aggregated items’ tags. Here’s the tag cloud from the current prototype site:
Just seeing that aggregate cloud makes me smile. I’ll have to work on things like adding a group-only tag cloud, and maybe a tag with date parameters (which could be REALLY useful to build a movie displaying the shifts in tag weights over the course of a semester or year…)
As an aside, I’m pretty sure that this is the first post that I’ve added to all of the main categories of my blog: General, Work, and Fun. I’m pretty sure there’s something to that…
“a website that aggregates all feeds published by students in a class/department/institution, and helps contextualize them in the various groups/cohorts/courses each student participates in” – has that been written down before like that? I think that’s the most coherent statement of “Eduglu” that I’ve seen to date (and the same way I understand it).
Hello, D’Arcy,
Yeah, there’s definitely something special about the tag cloud — all that information, just a click away — Just for kicks, I set the tagadelic page here: http://feeds.educon20.org/tagadelic to show 3000 tags — partly an experiment to see how the site would hold up under resource-expensive pageloads, but also because I like how it looked — it’s fun to navigate using FF’s Find in this page —
RE: “There are issues, to be sure, mostly with respect to honouring the original tags in the aggregated items” delicious seems to be the main culprit here — are you seeing this from any other site? I’m wondering how difficult it would be to set up special handling for delicious feeds. Given that delicious uses spaces, and not commas, to separate tags, any feed from delicious would need to be processed differently.
I have also seen some issues with WP sites sending poorly formed xml, and with tags getting munged as a result.
FWIW, the feeds.educon.org site is aggregating a few hundred feeds, and has over 10,000 nodes in just under a week. I set it up on a pretty anemic server to see how it would fare, and I’ve had to adjust some admin settings to smooth out cron runs, but for the most part it has just worked, without the need to make any adjustments to the hosting environment.
Cheers,
Bill
@scott: not sure, but that may be the first time I wrote it quite like that… It’s how I’ve been framing it (in my head at least) for some time, but I don’t think I’ve described it quite so succinctly before…
@bill: yeah, it’s del.icio.us and some rogue feeds that seem to be gumming up the works, so I resorted to the Yahoo! autotagging thing, which is currently clobbering all incoming tags anyway. I’ll try to make some sense of it tonight…
Mythical EduGliu, yum!
And what is all that stuff tagged “arcy”?
great stuff!
Leave it the Calgary Kid to make it happen, bravo!
Hey D’arcy,
Good post I think I implemented the same thing(tagedelic??) on http://hikealberta.com/hike/tags using Drupal and it works great.
Hoping that the same thing can be done in a WordPress MU environment which will be a required feature on a project I am working on at UBC.
-scott
Ever thought about adding in mail from listservs as nodes? That might create a lot of noise, but it might also round out an fascinating knowledge base.
http://drupal.org/project/mailhandler
You could probably use Yahoo’s autoclassification engine to create tags for the mail-based content via cron too.
Ahh, scope creep … gotta love it 😉