The Tag That Built Podcasting

An Accident of Plumbing That Became an Industry

In early 2001, Dave Winer added three attributes to an RSS item element and, without intending to, laid the groundwork for a distribution medium that would eventually attract billions of dollars in investment, prompt Spotify to spend hundreds of millions acquiring studios, and draw every major platform from Apple to YouTube into a format fight. The element was called enclosure. Its specification, even today, fits in a single sentence.

The enclosure tag takes exactly three attributes: url, length, and type. The url points to the media file. The length declares its size in bytes. The type specifies the MIME type — audio/mpeg for an MP3, for instance. That is the entirety of what the specification mandates. Nothing in the tag describes chapters, transcripts, artwork, or episode numbers. Nothing encodes a publishing schedule or a show name. The element simply tells a feed reader that a file exists somewhere and how large it is — enough for a client to decide whether to download the thing before the user asks for it.

Winer shipped the enclosure element as part of RSS 2.0, formalised in a specification he released in 2002. The idea had appeared in earlier thinking about the format, but the specific tag arrived in that period of RSS development when Winer was working to make the format useful beyond text headlines. The goal at the time was more prosaic than podcasting: the ability to attach any file — software updates, images, audio — to a feed item. The MP3 use case was one among several. History selected it above all others.

Adam Curry and the Receiver That Made It Real

The enclosure tag remained a curiosity until Adam Curry, the former MTV video jockey turned internet entrepreneur, began experimenting with it in 2000 and 2001. Curry was interested in whether RSS could be used to distribute audio files automatically — the listener's machine fetching overnight what would otherwise require a conscious visit to a download page. He collaborated with Winer on demonstrating the concept and later, in 2003 and 2004, wrote an AppleScript application called iPodder that automated the download of enclosure-attached audio files and synced them to an iPod. The name was clumsy; the concept was exact.

iPodder — later renamed Juice, after Apple's legal team took an interest in the prefix — gave the enclosure tag its missing half. The tag told a feed reader what file existed. iPodder told a computer what to do with that information: fetch it silently, on a schedule, without user intervention. The combination of a publication mechanism and a receiver that understood the convention is what made podcasting a distribution system rather than a novelty.

Curry released an early episode of what would become "The Daily Source Code" in August 2004, generally cited as one of the first programmes made deliberately for this mode of distribution. The word "podcast" itself appeared in print that same year, in an article by journalist Ben Hammersley in The Guardian, though the term's precise coinage is disputed. What is not disputed is that the enclosure tag was the enabling infrastructure beneath all of it.

A feed is a text file sitting on a server. Between the publisher and the reader there is a fetch, and nothing else.Photo: Tima Miroshnichenko / Pexels

What the Tag Specifies and What It Leaves to Convention

The minimalism of the enclosure element is both its strength and the explanation for two decades of supplementary invention. Because the tag says nothing about show-level metadata, the industry improvised. Apple's iTunes directory, launched in 2005, required producers to use a suite of iTunes-specific RSS extensions — <itunes:author>, <itunes:image>, <itunes:category> — to populate its database. Apple did not design those extensions collaboratively with the feed community; it published them as requirements for inclusion in what quickly became the dominant discovery mechanism. Producers adopted them not because the tags were technically elegant but because Apple's directory was where listeners looked.

The RSS specification's silence on episode numbering meant that no standard ordering mechanism existed. The RSS <guid> element identified individual items, and publication date could imply sequence, but nothing in the core format enforced it. <itunes:episode> and <itunes:season> arrived much later, in 2017, as Apple updated its namespace ahead of its next generation of podcast apps. By that point, the ecosystem had spent more than a decade relying on publication timestamp as a proxy for order.

The word "podcast" itself appeared in print that same year, in an article by journalist Ben Hammersley in The Guardian, though the term's precise coinage is disputed.

The size requirement embedded in the length attribute was practical in 2001, when dial-up connections made pre-flight knowledge of a download's weight genuinely useful. In a world of mobile streaming, it is largely vestigial — a number that feed producers still supply and clients largely ignore. Validation tools flag its absence as an error because the specification demands it, while modern players stream directly from the url without consulting length at all.

This gap between what the tag specifies and what the industry needs is precisely what the Podcasting 2.0 namespace project, led by Dave Jones and Adam Curry, has been trying to close since 2020. The namespace adds <podcast:transcript>, <podcast:chapters>, <podcast:soundbite>, <podcast:funding>, and the <podcast:value> tag for streaming micropayments via the Lightning Network. These extensions live in their own XML namespace, technically independent of the original enclosure element, but they are parasitic on it in the ecosystem sense: they only matter in a world where enclosure already exists and already works. The Podcasting 2.0 project's ambition is to keep podcast infrastructure genuinely open at a moment when Spotify, Apple, and YouTube each have platform incentives to replace open RSS distribution with proprietary systems.

The stakes are significant. Podcast advertising revenue in the United States now amounts to billions of dollars annually in recent measurements, according to IAB and PricewaterhouseCoopers tracking. The feed format underpinning the distribution of essentially all of that inventory is still RSS, still anchored to a three-attribute tag written in 2001 by one person, without a standards body, without a working group, without an RFC. The IETF never touched it. The W3C never standardised it. It became infrastructure by adoption alone.

That is both the miracle and the fragility of the enclosure tag. No institution protects it. No governance body can update it. What keeps it central to a multi-billion-dollar industry is the inertia of the receiver ecosystem — tens of thousands of apps, platforms, and directories that all know what to do when they see url, length, and type in an RSS feed item. Convention, in this case, has proven more durable than any standard.