The Evolution of RSS

[next]

The Evolution of RSS

By Andrew King ([email protected])

Abstract

We look at how RSS has evolved from its humble beginnings through present day and beyond. We survey all versions of RSS, including a feature comparison, a new RSS usage survey, and format and validation information. We also interview the people and explore the standards behind RSS. Learn how the newest version of RSS will move us towards a more Semantic Web.

Update (051401): Now includes more background standards material, 0.92 information, a Dan Libby interview, and extensive reference links.

Introduction

Rich Site Summary (RSS) is a lightweight XML vocabulary for describing metadata about Web sites, ideal for news syndication. Originated by UserLand Software in 1997 and used by Netscape to populate Netscape's My Netscape portal with external newsfeeds ("channels") RSS has taken on a life of its own and has become perhaps the most popular XML format today. Thousands of Web sites today use RSS as a "what's new" mechanism to drive traffic their way. This article discusses the various versions of RSS, the standards behind them (DTDs, RDF, namespaces), and the future of RSS. This article assumes you have some familiarity with RSS. For more information see our introduction to RSS.

Netscape Drops RSS

In the last week of April 2001 as part of their My Netscape redesign, AOL stopped including external RSS news feeds in their service. They also (probably inadvertently) removed the RSS 0.91 validator and most importantly the DTD, which most 0.91 RSS feeds point to. Thankfully, in response to feedback, Dan Libby (the author of the RSS 0.9/0.91 DTDs) reports that Netscape has replaced the missing files, though not the validator. While probably an oversight, the missing DTD caused some validating XML parsers to choke on the missing file, and the RSS community is abuzz. They ask: Where do we go from here?

This episode raises two important issues:

• Relying on dot-coms to house key public resources, and
• XML's use of single DTD locations that can cause single points of failure

The temporary loss of the USENET archives after Google's acquisition of Deja.com is a similar example.

Tristan Louis, freelance reporter at tnl.net says, "The problem with Netscape is showing some clear vulnerability in the use of XML as it relies on everyone pointing to a single DTD. When that DTD is gone, problems ensue.... We need to ensure redundancy across the network as a whole."

Rael Dornfest creator of O'Reilly's Meerkat RSS news portal and co-editor of the RSS 1.0 spec says, "RSS needs to, 1.0 or otherwise, move away from the static DTD model and reliance on a single point of availability (read: inevitable failure)."

Lars Marius Garshol, who wrote the original 0.9 DTD, says: "What I don't understand is all this fuss over Netscape removing the DTD. A well-designed RSS tool, whether it validates or not, would not use the DTD at Netscape's site in any case. There are several mechanisms which can be used to control the dereferencing of references from XML documents to their DTDs. These should be used, if not the result will be as described in this article."

RSS Lives

All is not lost, fellow syndicators. Your RSS files didn't break, and there are alternative RSS validators. Most of the aggregators (like Meerkat and My UserLand) work without a DTD as they have validation turn off or work with local copies. The good news is that Netscape put the RSS files back up, including Dan Libby's old Netscape RSS 0.91 spec, explaining their philosophy for RSS. Dan also reports Netscape posted his "futures" document, which "was actually the first RSS format I came up with," and has some similarities to RSS 1.0. Dan credits UserLand, appropriately, for their contribution to RSS 0.91.

https://my.netscape.com/publish/formats/rss-0.9.dtd
https://my.netscape.com/publish/formats/rss-0.91.dtd
https://my.netscape.com/publish/formats/schema-1.0.dtd
https://my.netscape.com/publish/formats/rss-spec-0.91.html
https://my.netscape.com/publish/help/futures.html

To find out more, check out the three RSS-related discussion lists at the end of this article.

RSS Validators

I asked Dave Winer (CEO of UserLand, primary contributor to the 0.91 spec, and author of the 0.92 spec) about validators now that Netscape/AOL/Time-Warner has removed theirs. It turns out that UserLand already has one. You just may not have known it. It's located at:

https://aggregator.userland.com

Dave updated his validator to accept 0.91, 0.92, and 1.0 RSS files. Just enter the URL of your RSS file and the validator will do its thing. Valid feeds can be registered into UserLand's public RSS database, which is used by some aggregators. UserLand's RSS validation/registration procedure is actually better than Netscape's now defunct process since it occurs in two easy steps.

Leigh Dodds has a prototype open source RSS 1.0 validator available at:

https://www.ldodds.com/rss_validator/1.0/validator.html

Leigh says that updating his script (based on Schematron, which uses XSLT style sheets) to validate 0.91 files would be easy, just plug in another XSLT.

Contents

• The Standards and People behind RSS
• RSS 0.9
• RSS 0.91
• RSS 0.92
• RSS 1.0
• RSS 0.91+ vs. RSS 1.0
• RSS Feature Summary, Usage Survey, Applications
• The Future and Reference

[next]