By Stephen Downes
Rich Site Summary, or RSS, is an XML format originally designed to list the changing contents of a news Website. Originally released by Netscape in 1997, it was used to allow readers to customize their personal Netscape Webpage and to create content indices inside the Netscape browser. Abandoned by Netscape, the format lived on through a grassroots content syndication movement and has evolved through several versions. Today, RSS is widely used by news Websites and Weblog authors.
Simplicity leads to popularity
The major attraction of RSS for Web developers is its simplicity. (In fact, RSS is often known by an alternative name, Really Simple Syndication). An RSS file can be created from scratch using nothing more than a simple text editor and sample file as a template, posted to a Web server as though it were a Webpage, and retrieved and read by a wide variety of applications. Additionally, and this probably accounts for its recent burst of popularity, numerous content management tools now create RSS files automatically and applications called headline readers enable users to view the contents.
A single RSS file, commonly known as a feed, consists of two major types of elements:
The channel element describes the Website title and base URL. Each item element contains an item title, URL, and a short description. Optional elements in an RSS file include a channel image and a form submission description. (See a sample RSS file at http://www.downes.ca/news/OLDaily.rss. )
Though originally designed for news content, and currently widely used by the news media, the most common use of RSS is to list Weblog contents. This became possible when the major blogging software applications—specifically, Blogger, Moveable Type, and Userland Radio—began to create RSS files automatically when bloggers entered and published a new submission. The structure of an RSS file mirrors the structure of a blog entry, so the user need not enter any additional information over and above the blog entry.
An RSS file will typically display the most recent content of a Website, usually ten items or so, updated whenever a new item is added. RSS files are therefore read on a regular basis by software applications known as harvesters or aggregators that scan for new entries and retrieve the data. An aggregator will check a large number of individual RSS files, returning to a given site once an hour or so. Consequently, when new material is published to a news site or Weblog, it is very quickly picked up and distributed.
Because RSS files are structured data, and because they are updated so frequently, they support content syndication much more easily than a Webpage. Popular aggregators, such as Blogdex, Daypop, Popdex, Technorati, Blogstreet and Feedster, are able to represent new content in a more user-friendly format than a regular search engine, including Yahoo! or Google. Because Weblog entries and news stories link to each other, these aggregators can quickly find the most popular new items. For many readers, a site such as the DayPop Top 40 is as useful a source of news as any online newspaper or portal—it’s far more current and not reflective of any editorial influence or control.
Though most readers use RSS by turning to an aggregator Website, many others use applications known as headline readers. A headline reader performs the same function as an aggregator, but is a stand-alone application that usually resides on the readers own computer (though some, such as Bloglines, are stand-alone Websites). Desktop readers, such as AmphetaDesk, FeedDemon and NewsGator, divide the screen into three panes:
For readers, the most commonly expressed benefit is convenience. RSS headline readers automatically flag new items, so readers need not search through a number of Websites looking for new content. Additionally, content is displayed first as a summary description, allowing readers to browse quickly through numerous items. RSS readers also provide readers with more choice and control because they can determine whether or not to subscribe to a given feed. And unlike email newsletters, which RSS feeds most resemble, the feeds do not contain spam or viruses.
The benefits of RSS have not been lost on educational technologists, with the result that some early work has been done to adapt the format to educational use. In their widely regarded paper and presentation, What’s the Fuss, Alan Levine, Brian Lamb and D’Arcy Norman demonstrate the use of RSS and a feature called trackback to facilitate the distribution of learning resources to novel audiences. Trackback allows the owner of a resource to know when it has been linked to by another user, and thus helps in the propagation of learning resources through a potential audience.
RSS is also being used to support the use of Weblogs in the classroom. In the weeks preceding this article, for example, staff and students at Centre d'Apprentissage du Haut-Madawaska posted 538 public and private blog entries among them. Rather than search each student’s page individually, a teacher or administrator simply uses an RSS aggregator to capture and display the day’s most recent posts. (See http://cahm.elg.ca.)
Additionally, an RSS aggregator can be used to create a specialized community of interest. The first example of this is my own Edu_RSS, which collects about 300 feeds related to learning technology and displays them in a single location, updating the results hourly. For educational technologists with more specialized interests, Edu_RSS also organizes the incoming items into a set of about 100 specific topics. Each topic generates its own RSS feed, so a person can keep track of all developments in the field of, say, learning objects, by subscribing to this single feed. (See Edu_RSS at http://www.downes.ca/cgi-bin/edu_rss/edu_rss.cgi.)
The operators of learning object repositories have also started to experiment with RSS feeds listing recently added or topical learning resources; collections providing feeds now include the Maricopa Learning Exchange, Merlot, EdNA, CAREO, and the UK Centre for Materials Education. These feeds may be read in any RSS headline reader. In addition, such software as the Distributed Learning Object Repository Network (DLORN) harvests the feeds from these repositories and displays the results in a centrally accessible site, which greatly eases the search for learning objects from a wide range of sources.
Flexibility leads to compatibility
Operating parallel to RSS, and using a slightly different format, the Open Archives Initiative (OAI) operates on the same principle. Instead of using RSS, OAI feeds list resources using (typically) Dublin Core, which while providing the same type of information as may be found in an RSS channel, offers more detailed information about authorship and publication data. An OAI site typically requires the installation of an OAI server, which in addition to supporting plain harvesting allows for a site-specific search (though recently OAI has released a harvest-only version of the format.
The OAI initiative has been widely embraced by the academic community and has supported several spin-offs, the most notable being MIT’s DSpace open archiving service. The Institutional Archives Registry now lists about 180 feeds containing many thousands of academic articles. Another aggregation service, OAIster reports as of this writing to have collected 3,063,884 records from 277 institutions.
It is only a matter of time before the RSS and OAI worlds merge. Authors of RSS software are by now used to working with different formats. In the seven years since the format was released, there have been nine different flavors of RSS, the most popular being the original RSS 0.91 format used by Netscape, the RDF-flavoured RSS 1.0 format, and the most recent RSS 2.0 format. An alternative format, called Atom, is now being supported by the major blogging software suppliers. And though not widely used, specialized formats such as NewsML are being used by specific communities.
Most aggregators and headline readers are indifferent to the original format provided by an RSS feed. Transformations between types of XML may be accomplished by software engines or through the use of XSLT files. These transformation files make the different types of RSS and similar formats essentially interchangeable. Regardless of what version the original RSS feed uses, the result looks more or less the same when displayed to a reader. Indeed, one of the reasons RSS has become so popular is that for the vast majority of content producers and consumers, the RSS encoding remains completely behind the scenes, used only by the applications to communicate with each other.
RSS also supports compatibility with additional XML formats through the use of extensions or modules. (See http://web.resource.org/rss/1.0/modules.) These are formally supported by RSS 1.0 and RSS 2.0, but they are sometimes used in other formats as well. Major extensions include Dublic Core, Syndication, and Content modules, allowing for a more precise description of individual items. Another widely used, though unofficial, module is the Creative Commons module, which allows a feed provider to refer to a Creative Commons license from within the RSS feed. Additional proposed modules include annotation, DC Terms, event, and email. (See http://web.resource.org/rss/1.0/modules/proposed.html.)
The more detailed description of learning resources may also be supported using RSS modules. Though not formally accepted or implemented, my own RSS_LOM is an RSS 1.0 module designed to allow the inclusion of IEEE-LOM metadata in an RSS file. This allows RSS feed items to include such learning-specific metadata as TypicalAgeRange or InteractionLevel. The use of modules in RSS files describing learning resources will allow resource owners and users to add a wide range of information not envisioned in the original LOM specification, including evaluation and peer review information, event metadata, and digital rights information. (See http://www.downes.ca/files/resource_profiles.htm.)
Another change likely to spread through the RSS world over the next few months is the integration of social networking metadata with RSS content metadata. The popularity of such social networking sites as Friendster and Orkut has shown that there is a need for individuals to describe themselves and their relations with other individuals. These descriptions, most commonly found in a Friend of a Friend (FOAF) file, may be referred to by an RSS file using a social networking module. A FOAF file is another form of XML file, and so may be created read by many of the same tools now creating and reading RSS files. Adding social networking information to RSS allows for even more finely grained filtering and searching, as author information may now also be included in the search parameters. (See http://www.downes.ca/cgi-bin/Website/view.cgi?dbs=Article&key=1076791198.)
The future leads to…
To be sure, RSS will evolve rapidly over the next few years.
It’s poised to be exposed to a great deal of rhetoric,
and is on the verge of being widely commercialized, with the
inevitable cycle of hype and disappointment that will follow.
That said, RSS is a technology with a strong future, strong
because of its simplicity, flexibility, and utility. Although
RSS is not the semantic Web originally dreamed of in the laboratory,
with finely grained and standardized element descriptions and
canonical vocabularies, it is a technology that has proved itself,
and evolved roughshod, though the much grittier practice of
grassroots development. There is, I think, a lesson in that.
Stephen Downes is senior research officer for the Institute for Information Technology, National Research Council, Canada. Downes specializes in research in online learning, online communities, and knowledge management. He also publishes a daily newsletter (OLDaily) about online learning.
He can be contacted at [email protected].