Introduction
Since the 1990s, the World Wide Web has mainly served as a platform for publishing documents meant for human eyes. These documents are typically written in HTML, which offers only a basic layer of structure—like indicating where a paragraph begins or which word should be emphasized. Add a dash of CSS, and you can dress up that structure with visual flair, such as making all paragraphs appear in tiny gray sans-serif text. This might make you look trendy, but it also risks alienating older readers who struggle to read such fine print. That’s essentially the extent of “structure” on the web as we know it.
The Structure Gap
Consider a simple example: you mention a book on a webpage. You might write something like:
Goodnight Moon by Margaret Wise Brown
Illustrated by Clement Hurd
Harper & Brothers, 1947
ISBN 0-06-443017-0
To a human, it’s clear that this is a book citation. But a naive computer program scanning the page would see little more than a jumble of text. The only hint of structure is the bold formatting on the title. There’s no machine-readable indication that this is a book, let alone details like the author, illustrator, publisher, or ISBN.
The Semantic Web Vision
As early as 1999, visionaries like Tim Berners-Lee foresaw a more intelligent web. In his book Weaving the Web, he wrote:
“I have a dream for the Web [in which computers] become capable of analyzing all the data on the Web – the content, links, and transactions between people and computers. A ‘Semantic Web’, which makes this possible, has yet to emerge, but when it does, the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines. The ‘intelligent agents’ people have touted for ages will finally materialize.”
This vision of the Semantic Web promised to give web content much richer, machine-readable meaning. For instance, instead of just bolding a book title, you could mark it up using standards like those from schema.org. You might employ formats such as RDF or JSON-LD to embed structured data directly into your HTML, explicitly telling computers, “Hey, this is a book!”
The Adoption Challenge
In theory, this sounds great. In practice, it’s been a hard sell. Adding semantic markup requires extra effort—it feels like homework on top of writing a blog post. After carefully crafting your content to be human-readable, it’s tough to muster the mental energy to also make it computer-readable. And unless there’s already a computer program actively consuming that data, you might wonder why you should bother. As a result, despite being around since the late 1990s, semantic markup remains rare on the actual web.
The Path Forward
We believe this situation can—and must—change. Human progress depends on making information increasingly accessible, not just to humans but also to AI systems and traditional computer programs. The key insight is simple: people will only add semantic markup to their web pages if doing so is easy and immediately beneficial. If we can lower the barrier to entry, we can unlock the web’s full potential as a structured, machine-readable library of knowledge.
Efforts like the Block Protocol are now working toward this goal, aiming to make structured data a natural and painless part of web publishing. By providing reusable, interactive blocks that carry their own semantic meaning, they hope to fulfill Berners-Lee’s dream—one where the web is just as useful for machines as it is for people.