Computing, IT & AI

RSS, RDF and Schema.org

1997-2011 CERamanathan V. GuhaRSS 0.9 (Netscape, 1999); Schema.org (2011)Diaspora (work abroad)

1997

Web standards that let machines read the structure of web pages: feeds, linked data and structured search results.

Dated: 1997–2011
As scholars put it: 1997-2011 CE
Originator: Ramanathan V. Guha
Source text: RSS 0.9 (Netscape, 1999); Schema.org (2011)
Region: Diaspora (work abroad)
Discipline: Computing, IT & AI
Evidence: Documented
Confidence: High
Verification: Verified-web

A knowledge engineer looks at the web

Ramanathan V. Guha studied mechanical engineering at IIT Madras, then went to Berkeley and Stanford, where he finished a PhD on formalising context in 1992 ¹. From 1987 to 1994 he co-led Cyc with Douglas Lenat, a project to write down common-sense knowledge in a form a computer could reason with, and designed much of its language, CycL ¹.

He brought that habit of mind to the early web. Joining Apple in 1994, he saw web pages as text with no description of what they were about ⁵. His answer was the Meta Content Framework (MCF), a format "to provide an adequate system for representing a wide range of information about content" ⁵. Apple used it in HotSauce, a viewer that let you fly through a site's structure in 3D, and several thousand sites, Yahoo among them, published MCF files before Steve Jobs cut the project on his return ⁵.

From MCF to RDF to RSS at Netscape

Guha moved to Netscape in 1997 and was paired with Tim Bray to recast MCF in XML ¹⁵. Their W3C submission of June 1997 modelled metadata as a directed labelled graph with a small, self-describing type system ⁷. That idea became the core of the Resource Description Framework, RDF, the W3C standard for linked data ¹⁵.

Netscape then needed a way for sites to describe their latest headlines for its My Netscape portal ⁵. RSS 0.90, published by Netscape on 15 March 1999, was authored by Guha and built on RDF ⁵⁶. Colleagues found RDF too heavy for ordinary publishers, and Dan Libby wrote a simpler, RDF-free RSS 0.91 in July 1999 ⁵⁶.

The feed wars and the podcast

RSS then split. Dave Winer's UserLand Software issued its own 0.91 in June 2000, then 0.92 and, in August 2002, RSS 2.0 ⁶. A working group that included Guha went the other way and kept RDF at the centre of RSS 1.0 in 2000 ⁵. RSS 2.0 won in practice; its specification passed to Harvard's Berkman Center in 2003 and is now maintained by the RSS Advisory Board ⁶.

The feed format outlived the blog boom because podcasting was built on it. Apple Podcasts, for example, requires that any podcast feed submitted to it follow the RSS 2.0 specification ¹⁰. A podcast app is, at heart, an RSS reader.

Schema.org: one vocabulary for every search engine

After Epinions, a startup and IBM Research, Guha joined Google as a Google Fellow in 2005 ¹. On 2 June 2011 he announced Schema.org, a joint project of Google, Bing and Yahoo "to create and support a common vocabulary for structured data markup on web pages" ⁸. Google had shown rich snippets for reviews and people since 2009, but publishers had to guess which markup each engine would read; Schema.org added more than a hundred categories at launch, from movies and music to products and places ⁸. Yandex joined, and since 2015 the vocabulary has been developed in an open W3C Community Group, with Guha chairing the Steering Group ¹¹.

Adoption was fast. By late 2015, in a sample of 10 billion pages, 31.3 percent carried Schema.org markup, up from 22 percent a year earlier ⁹. The same markup feeds Google's Knowledge Graph panels, recipe carousels and Gmail's structured emails ⁹. In 2015 the ACM made Guha a Fellow "for contributions to structured data representation and specification and their impact on the Web" ¹.

Background

R. V. Guha, an IIT Madras graduate, worked on knowledge representation (Cyc and the Meta Content Framework at Apple). At Netscape he contributed to RDF and created RSS in 1999 with Dan Libby. At Google he started Schema.org in 2011, a shared vocabulary that the major search engines agreed to read.

Significance

RSS made blogs and podcasts subscribable. Schema.org markup is on a large share of web pages and powers rich search results.

Path to modern innovation

Every podcast is distributed as an RSS feed.

Global impact

Podcast distribution.
Rich results and knowledge panels in search engines.

Primary sources

4Schema.org
7Meta Content Framework Using XML (W3C Note) R. V. Guha and Tim Bray, Netscape Communications / W3C, 1997. The 1997 submission that modelled metadata as a labelled graph, a direct ancestor of RDF