April 22, 2007

An MIT OpenCourseWare Course via an OPML Feed

Over the weekend, I noticed that MIT OpenCourseWare courses were offering a "Download this Course" option. Intrigued, I immediately grabbed a copy of the CMS.610 / CMS.922 Media Industries and Systems, Spring 2006 course and had a poke around inside it.

The download bundle is - I guess(?!) - a standard (?) IMS - err - package? (can you tell I'm not up on educational material interoperability standards?!;-)

That is to say, the zip archive file opens into a set of nested directories with an imsmanifest.xml document. The archive file also contained copies of the HTML pages used on the course website as well as the PDF versions of the course lecture notes.

What I had been hoping for was a 'clean' XML version of the course webpages (i.e. a single source document from which they had been generated). Some hope!

Anyway, my initial enthusiasm curbed somewhat, I felt there was still some mileage to be made in RSSifiying the course, for three main reasons.

Firstly to see if what came out was "sensible" - i.e. whether a feed based version of the course could be sensibly viewed in a Grazr or Stringle environment, for example.

Secondly, to get a feeling for how to decompose a feed-delivered version of the course into separate component feeds.

Thirdly, to see whether those component feeds could be automatically generated from the IMS manifest, or scraped from the MIT OpenCourseware web pages.

The course - as published on the MIT OpenCourseWare website - is structured according to the course syllabus, readings, lecture notes, projects (i.e. group and individual activities) and related resources.


As a first pass, I have generated an OPML version of the course that follows a similar structure. At the moment, this OPML feed is - to all intents and purposes - a monolithic feed. Over the coming weeks I intend to disaggregate it into separate component feeds that can be reaggregated in an OPML feed that pulls in separate OPML and RSS component feeds.

You can view the OPML version of the course here: MIT OCW CMS610 Spring, 2006, Comparative Media (via Grazr).


Note that at the current time, some of the links (to New York Times) resources require a NYT login. I will patch these as and when I can. In addition, the links to audio-visual resources should really be made available via an embedded player, as well as link to the original site.

A cursory scan of the OpenCourseWare site suggests that the way individual web pages are composed across different courses is rather arbitrary, which will make scraping the site to automate feed production somewhat tricky.

I should be thankful for small mercies in the way OpenLearn content is bundled as XML I suppose!

PS this is spot on.

Posted by ajh59 at April 22, 2007 11:04 PM

Nice to see you tackling another Open CourseWare provider as well as OpenLearn!!

I had a quick look, and yes they are using a "standard IMS" way of sharing content. This one is the IMS Content Packaging (CP) standard. See http://www.imsglobal.org/content/packaging/index.html

You may be interested to know that the SCORM standard and the IMS new thing "Common Cartridge" both use a similar method (html files and an imsmanifest.xml) to share content. Common Cartridge (CC) hasn't been publicly released yet.

I've been involved recently in hand-crafting a few OpenLearn units into CC. Joel Greenberg presented them at the IMS Learning Impact Conference, and they were very well received! We (OpenLearn) will be looking at offering SCORM, IMS CP and IMS CC download formats later this year.

They are useful formats if you want to take content and stick it in another VLE, but I think your RSS feed from OU XML is more useful from an automatic scraping point of view. The drawback of sharing just HTML is that you completely loose the structure, which makes it a nightmare to do anything automatic with it.

Posted by: Jenny Gray at April 23, 2007 09:28 AM