I happened upon Middlesell early this morning (this site is an independent homepage for UMW students established by the inimitable Bobby Durette) only to discover that posts from around UMW Blogs are being fed into this space. How fascinating!
Now it really shouldn’t come as a surprise to me, for Bobby is a remarkable guy who I have had the pleasure of listening to and talking with on several occasions. Here’s a brief tech bio of one of UMW’s finest: he started Middlesell as a hand-coded PHP project in his Freshman year; during his Sophomore year–after a recommendation from Martha Burtis and Cathy Derecki– he decided to use Drupal as the site’s CMS; and now he is currently a Junior and has effectively developed a community site that is a regular online destination for a large part of the UMW campus. Most impressive, Jedi! What I particularly like about Bobby is that he is wise beyond his years and a consummate gentleman, and I say all this in spite of his using Drupal as his CMS of choice mind you 😉
So, when I saw that he was pulling posts from the UMW Blogs sitewide feed into Middlesell I really shouldn’t have been all that surprised. He’s a smart guy and recognizes the value of solid content that is directly targeted to his audience, not to mention that he is intimately familiar with with the wondrous power of RSS aggregation. All of which makes me think about a thread on a post D’Arcy Norman wrote recently about creating a blog directory for WordPress Multi-User wherein Bill Fitzgerald suggested how Drupal could be easily configured with views to offer this functionality, but Drupal disciples always mention the Views Module when they want to discount WordPress (and they usually mention the CCK module soon after this) so I wasn’t taking the bait.
What was most interesting to me, however, was something Mr. Fitzgerald (I’ll appeal to his sense of formality) describes in his second comment on this thread that amounts to a more sophisticated interaction between these two applications that would utilize WPMu as the publishing platform and would run its sitewide feed through Drupal to harness functionality such as searchable aggregation, mapped external tags, and OPML generation. Mr. Fitzgerald says the following in his comment on D’Arcy’s blog:
The more I think about it, though, the more I like the idea of an OPML feed coming out of a blog directory[in this case WPMu], and being fed into a searchable aggregator. Keep in mind that the feedapi allows for mapping of external tags into Drupalâ€™s core taxonomy, and that you can then slice those imported feeds up into new sections via views, and then expose that via an rss feedâ€¦
Yup. I taste some Drupal-Wordpress kool-aid, consumed straight out of the grail
So, if I understand this correctly, and please correct me if I don’t, WPMu remains the individualized publishing platform and Drupal handles the aggregation, searching , and display. Drupal wrapped around WPMu in other words (something the good folks running Arizona State University’s blogs mentioned to me at ELI last year, but I couldn’t fully grasp it).
So, Mr. Fitzgerald, do we have a shared project/vision finally?
And, Mr. Durette, might you be interested in experimenting with such a feeding and parsing of content from UMW Blogs into Middlesell? You are one of our local Drupal aces, and your feeding posts from UMW Blogs into Middlesell (I’m assuming this was you, by the way) really brought all these loose connections home to me.
If neither of you are interested, I’ll just have to take this silence as a tacit understanding that Drupal sounds great in blog comments and can feed posts into a site like any old aggregator, but just doesn’t have what it takes to push this baby to the next level. No sweat, I just won’t let you forget it for a while is all.
You got it, Rev! Students publish wherever they wish (WP, FB, whatever) and the Drupal site sucks it all together under the umbrella of the course/context/institution. This is one key component of the EduGlu stuff we’ve been talking about. THIS. IS. EDUGLU! 😉 But, in the full-on EduGlu, you’d be able to say “give me an OPML for all blogs posting stuff for Humanities 311 in the Fall 2008 semester” and it’d just do the funky stuff to pull together the list of URLs. Or, instead of an OPML, you could say “build me a web page displaying the most recent stuff on assignment #3 in HUM 311 FALL 2008” etc…
So, in your opinion D’Arcy, given the experience you have with Drupal, how closely can this process be approximated. Brian was right when he said this idea of eduglu was apparent in Planet NV and UMW Blogs (in the comments to this great post by him) on a broader scale, can Drupal handle more granular functionality as Bill suggests?
I think Drupal can handle much of it right out of the box. But not the full concept. I’ve had Drupal sites running as aggregators, happily sucking in feeds, making nodes of posts, and inheriting tags, etc…
but there’s more to it than simply aggregating. there’s also the dynamic directory (BlogBridge FeedLibrary is the closest thing I’ve seen on that front), and integration with a school’s student info. system (for classes, semesters, cohorts, groups, etc…) to make the process as seamless as possible. The prototypes work manually on a small scale, but for this to work, it’s got to scale to 10,000 or even 100,000 students.
I’m still struggling with how to put together the full meal deal – which is why I haven’t finished it yet 🙂 I’ve got a few sandbox apps that model various bits of the system, but for it to work, I need to figure out ways to tie them all together seamlessly and effortlessly…
Greetings, Mr. Good Reverend James,
First, I would like to say that, wrt “So, if I understand this correctly, and please correct me if I donâ€™t, WPMu remains the individualized publishing platform and Drupal handles the aggregation, searching , and display.” — absolutely. Except that as far as Drupal is concerned, it can be from any site on the web that generates a clean Atom or RSS feed.
I put together a proof of concept of sorts at http://feeds.openacademic.org — nothing big, but it starts to show how these pieces can fit together. This will start to address some of D’Arcy’s questions as well —
You have at least three ways of tagging every post that comes into the site (and, as Drupal allows you to have multiple taxonomies describing a single piece of content, you can get pretty flexible here):
1. Inherit the tags from the original post;
2. Create a set of terms that will be appended to every imported post on a feed (this can allow you to re-tag content with a descriptor that can, for example, streamline how/to whom the data gets routed);
3. (and this is my favorite) Map elements of a feed to specific taxonomies/fields — so, if a feed contains additional metadata about the content, you can map that to a specific field, and create another method of sorting/routing/displaying content.
Between 2 and 3, you start to approach the functionality that D’Arcy is describing. There will definitely be new code required to extend what we currently have in place, and particularly to improve the UI to make it EASY to use, but extending/creating UI tweaks is a far cry from having to build up from scratch.
Within the current proof of concept, you can filter by multiple taxonomy terms, by author, by date, and the site also generates feeds of all imported content. Most importantly, it also retains a link back the original post.
We’re getting closer. I’d love to see what could happen if we got some folks to get painfully specific about what would be needed to make this work.