Dave Winer and RSS

Abstract

Dave Winer did not invent blogging, but he did more than anyone else to build the infrastructure that made it viable at scale. He pioneered blog publishing tools, shaped RSS through its successive versions, added the enclosure element that accidentally enabled podcasting, and spent decades fighting for open standards against platform capture. He is also a combative figure who feuded with nearly every collaborator involved in building these systems, which has meant that the infrastructure he created is often better known than the name of the man who built it. RSS survived Google Reader, survived social media’s attempt to replace it, and as of the mid-2020s remains the invisible backbone of global podcast distribution.

Dave Winer. Image: Joi Ito from Inbamura, Japan, CC BY 2.0, via Wikimedia Commons.

UserLand and the Early Web

David Winer was born on May 2, 1955, in Queens, New York. He grew up in New York and New Orleans, studied mathematics at Tulane University, and pursued graduate work in computer science at the University of Wisconsin–Madison before leaving academia to build software. In 1988, he founded UserLand Software in California. UserLand’s core product was Frontier, a scripting language and development environment for the Macintosh, designed for power users who wanted to automate and extend the system beyond what Apple’s built-in tools provided.

Frontier competed in a narrow niche. It was well-regarded by its users, technically innovative, and commercially modest. What gave Winer influence beyond UserLand’s market position was his willingness to engage with the early public internet before it had obvious commercial applications, and his instinct to build infrastructure rather than applications, to make things easier for other people to build.

In November 1994, Winer began publishing DaveNet, a stream-of-consciousness essay series distributed by email and archived on the web. In February 1997 he launched Scripting News, a daily journal on his website. It covered software development, internet technology, and whatever he was thinking about that day. The word “blog” would not exist until 1999; what Winer was doing was writing a regularly updated personal website in a conversational register, with dated entries, in reverse chronological order. He was doing this before nearly anyone else, and he was doing it in public, under his own name, on his own server.

Blogging as Infrastructure

By 1997, a small community of people maintained regularly updated personal websites in the format that would eventually be called blogs. Winer recognized that the practice needed infrastructure three distinct layers: tools to make writing and publishing easier, formats to enable content to travel between systems, and aggregators to allow readers to follow many sources at once.

UserLand shipped Manila in 1999 and Radio UserLand in 2002, early blog hosting and authoring tools that allowed non-programmers to maintain regularly updated websites. These preceded Blogger (Evan Williams, 1999, acquired by Google in 2003) and were used by the first generation of technology bloggers, early journalists experimenting with personal publishing, and organizations experimenting with institutional blogging.

More technically consequential was Winer’s work on XML-RPC, a lightweight remote procedure call protocol using XML over HTTP, developed with Microsoft in 1998. XML-RPC enabled blogging tools from different vendors to communicate with each other and with hosting platforms. It became the basis for WordPress’s first external API and remained the standard for blog tool interoperability for years. It was an example of the pattern that would define Winer’s career: building the infrastructure layer that makes other people’s tools interoperate, rather than the application that users see directly.

RSS: The Feed Format Wars

The history of RSS is genuinely contested, and Winer was central to the most important controversies.

In March 1999, Netscape developed RSS 0.90, technically an application of the RDF (Resource Description Framework) format that Tim Berners-Lee was promoting as the foundation of the Semantic Web. Netscape wanted to pull content from third-party websites into its My Netscape portal; RSS 0.90 was the format that made this possible. The RDF basis made RSS 0.90 semantically rich but technically demanding to implement. RSS also delivered, cheaply, on the promise that “push” technology had made and broken two years earlier: the 1996–97 hype around PointCast had tried to send news to users automatically, but with a heavyweight proprietary client that choked networks. A lightweight XML feed did the same job without the bandwidth bill.

In July 1999, Netscape’s Dan Libby published RSS 0.91, a simplified version that dropped the RDF framework entirely, replacing it with a straightforward XML format that any developer could implement in an afternoon. The revision folded in elements from the scriptingNews feed format Winer had developed independently and run on his own site since December 1997. The simplification was decisive: within months, RSS 0.91 was far more widely implemented than its predecessor.

When Netscape abandoned the RSS project (the company was disintegrating under the pressure of the browser wars with Microsoft) Winer became the de facto steward of the standard, republishing the RSS 0.91 specification on UserLand’s site in June 2000. He published RSS 0.92 in December 2000, adding the <enclosure> element and categories. He published RSS 2.0 in September 2002, making backward compatibility a firm commitment and closing the specification against further revision. In July 2003, he transferred the RSS 2.0 specification to the Berkman Center for Internet & Society at Harvard University (since 2016 the Berkman Klein Center) under a Creative Commons license, making its governance transparent and its permanence assured.

Info

The <enclosure> element (introduced in RSS 0.92 and carried into RSS 2.0) allowed feed items to attach binary files (typically audio files) to entries alongside the text content. Winer began experimenting with it in January 2001, famously enclosing a Grateful Dead song in his blog’s feed. The element was designed for any binary file, but it was audio that found immediate use. Adam Curry (a former MTV VJ who had become a technology entrepreneur, and who had pushed Winer for exactly this capability) released iPodder in 2004, aggregator software that downloaded enclosures automatically and synchronized them to portable music players. The combination of the RSS <enclosure> element, the iPodder aggregator, and Apple’s iPod was podcasting. No single company controlled the infrastructure. No approval process was required. Anyone could produce a podcast by maintaining an RSS feed with audio attachments.

A competing group of developers, frustrated by Winer’s unilateral control of the RSS specification and by what they described as his combative and unresponsive engagement with feedback, created Atom in 2003. Atom was developed through the IETF standards process, open, documented, formally governed. Its technical design was cleaner than RSS 2.0 in several respects: better internationalization, clearer separation of content types, more consistent date handling. Google adopted Atom for its early APIs. Many blog platforms supported both formats.

The format split never resolved into a single winner. RSS 2.0 dominated for podcasts; Atom was more common in certain web APIs. The competition was not obviously harmful (both formats remained readable by every major aggregator) but it demonstrated the cost of Winer’s governance style: by refusing to allow community-driven evolution of RSS, he had pushed frustrated engineers to build a parallel standard rather than improve the existing one.

Podcasting and the Open Infrastructure

The term “podcast” was coined by journalist Ben Hammersley in a Guardian column on February 12, 2004. Hammersley noted, in passing, that cheap microphones, RSS enclosures, and the iPod’s popularity had created the technical conditions for a “new boom in amateur radio.” The term stuck and spread faster than the practice it described; within a year, “podcast” was being used by mainstream media to describe a new form of audio publishing.

What distinguished podcasting from every other media format of the first internet era was its radical openness. A podcast was an RSS feed with audio attachments, hosted on any web server, indexed by aggregators that anyone could build, subscribable by any software that could parse RSS. There was no central registry, no required platform, no approval process, no revenue share, no exclusivity requirement. This architecture reflected Winer’s deliberate philosophy: the infrastructure should belong to no one, so that no one can be excluded from it.

Apple added a podcast directory to iTunes in June 2005, which dramatically increased podcast discovery and accelerated growth. But Apple’s directory was a pointer to RSS feeds hosted on independent servers; it was not a hosting platform or a content gate. Adding a podcast to the iTunes directory required no approval beyond the directory’s own review process, and the audio files remained on the publisher’s own server.

The openness that Winer’s infrastructure provided would prove crucial twenty years later.

Dead End: Platform Enclosure

Warning

Beginning around 2019, major platforms attempted to enclose the open podcasting ecosystem behind proprietary walls. Spotify acquired Gimlet Media for approximately $230 million, then Anchor (a podcast hosting platform), then Parcast, and signed an exclusive deal with Joe Rogan reportedly worth $100 million. The strategy was to make Spotify indispensable to podcast listening, to own discovery, distribution, and measurement in the way that Netflix owned streaming video and Spotify already owned music streaming. Apple responded by adding subscription features to its podcast infrastructure but kept the directory open and the RSS standard intact. The enclosure strategy failed on its own terms: Spotify never captured a majority of podcast listening, open RSS-based applications (Apple Podcasts, Overcast, Pocket Casts) kept a large share of the audience, and in 2024 Spotify gave up exclusivity for its biggest acquisition, signing a new Rogan deal that put the show back on Apple Podcasts and YouTube. The open infrastructure Winer helped design proved more resilient to platform capture than almost any other media distribution format of its era. The comparison with music, video, and social media (all of which consolidated rapidly around a handful of platforms) is striking. The absence of a central registry or a platform-controlled distribution layer meant there was nothing to buy, no chokepoint to acquire. Text-feed RSS proved more fragile than its podcast cousin: it had acquired a chokepoint after all in Google Reader, the client where most non-technical users read their feeds, and when Google shut Reader down in 2013 the mainstream audience for feed reading scattered and never reassembled.

Winer’s Legacy and the Question of Credit

Winer’s influence on web infrastructure is disproportionate to his public recognition. The blog as a form of publishing, the RSS feed as a distribution mechanism, the podcast as an open audio format, all of these things rest, in part, on technical work Winer did. The infrastructure is so deeply embedded in how the web works that its origins are invisible to most of its users.

Part of the reason for the recognition gap is character. Winer’s blog was, for many years, a place where he argued at length with collaborators, criticized journalists by name for misrepresenting his contributions, and conducted feuds that he documented in public detail. His dispute with Adam Curry over podcasting credit (Curry became publicly associated with podcasting in ways that minimized the technical infrastructure Winer had built) generated a prolonged, visible fight. His conflict with the Atom working group, whom he accused of acting in bad faith by building a competing standard rather than working with him, produced years of public antagonism. His feuds with Harvard’s Berkman Center over the governance of the RSS specification, after he had donated it to them, were documented in his blog.

These conflicts did not change the underlying facts of what Winer built. But they shaped how the story was told, and by whom. Journalists who had been attacked in Winer’s blog were not inclined to write sympathetic accounts of his contributions. Developers who had participated in the Atom process described Winer’s stewardship of RSS as a cautionary tale about open standards controlled by individuals.

Winer was a Fellow at Harvard’s Berkman Center for Internet & Society from 2003 to 2004, where he ran the Weblogs at Harvard Law project and organized the first BloggerCon conferences, promoting blogging as a tool for journalism and civic engagement. He has published Scripting News nearly every day since 1997, one of the longest continuously maintained blogs on the web. His position on open standards has been consistent throughout: no single company should own the feed format, no platform should control distribution, and the infrastructure of publishing should remain as open as the postal service.

The longevity of RSS (used daily by hundreds of millions of people through podcast apps, though most of them have never heard of RSS) is the best evidence that he was right about the infrastructure, whatever the disputes about the person.

The broader context of blogging and online publishing is covered in The JavaScript Revolution; the open standards movement in The Open Source Revolution.