What Alan Kay Got Right About Documents and the Web (And Why It Still Matters)
markdown formatted blog content
The Document That Never Was
There's a famous thought experiment in software design: What if browsers had stayed true to their document-viewing origins? No JavaScript frameworks, no SPAs, no real-time collaboration — just nicely formatted text with some images.
Alan Kay, the visionary behind Smalltalk and much of modern object-oriented programming, had a sharp response to this hypothetical: Actually quite the opposite.
His point wasn't that browsers should have remained simple. It was that we misunderstood what a "document" could be.
Beyond Static Paper
Kay argued that if we define "document" as merely an imitation of old static text media, we've already lost the game. But what if documents were alive? What if they could think, respond, and adapt?
This isn't abstract philosophy. It maps directly to technologies like:
- HyperCard (1987) — a "hypermedia" system where cards could contain code, graphics, and links
- Smalltalk — where everything, including UI elements, was an object with behavior
- NeWS and HyperLook — attempts at networked, interactive document systems
These weren't just experimental toys. They were early attempts at what Kay called "programming for the rest of us" — systems where end users could interact with and modify living documents.
Where We Landed
The modern web took a different path. We got HTML as a document format, CSS for styling, and JavaScript for interactivity bolted on separately. The separation was practical, but it created a divide: documents on one side, applications on the other.
But here's the thing — Kay wasn't necessarily wrong. He was just early.
Today, we're seeing his vision resurface in new forms:
- AI-integrated documents that understand context and respond intelligently
- Real-time collaborative tools where documents live and breathe with multiple authors
- Progressive web apps that blur the line between document and application
- Vibe coding tools that let developers express intent rather than write every line
The document of 2024 can do things HyperCard only dreamed of. It just took us 35 years to get there.
The Takeaway for Developers
Kay's insight is simple but profound: Don't design for the constraints of old media.
When you're building web applications today, ask yourself:
- Am I treating the browser as a document viewer with extra features?
- Or am I designing interactive experiences that happen to live on the web?
- Where can AI and modern tools let users do more than just consume content?
The best web experiences today — Figma, Notion, Linear, the tools we actually love using — aren't documents that do stuff. They're experiences that happen to support documents.
Final Thought
Alan Kay's ideas have a way of being right just before the technology catches up. If you're building for the web in 2024, you're not just choosing a platform. You're deciding what a "document" means for the next generation of users.
Choose wisely.
What do you think — should browsers have stayed simpler, or are we finally living in the document Kay imagined? Drop your thoughts below.
Read in other languages: