Emergence of the web
The Text Only No Code Web
In 1989 Sir Tim Berners-Lee builds the first killer app for the Internet: the web browser and the web server. You’re almost certainly using a derivation of that browser to read this article right now.
This is the first phrasing of the idea of what the web is. We see an applied philosophy, a thesis in code that focuses foremost on a human readable service; knowledge by humans for humans. It is as an online newspaper that allows anybody in the world to participate in a rhizomatic conversation at planetary scales.
This early implementation of web doesn't exist in a vacuum; it is predated and overlaps with many associated phenomena. Television is ubiquitous. People like Donna Harroway have already articulated an idea of merging human and machine (if perhaps tongue in cheek). Video games are becoming popular. Online games are emerging. There’s significant speculation and Science Fiction. Cyberspace has been proposed.
What the web gets right is that it is defined as a protocol (http). Since the Internet consists of several protocols we still often use http:// or https:// when specifying an URL. And browsers still support other protocols such as ftp:// but these are largely unused today.
Also what is unexpected in this early vision is a powerful idea of collective authorship, style and aesthetics. It allows lay persons to use extremely powerful tools to organize words into titles, chapters, the structure we expect from text - and it reduces that cost to zero. This accessible grammar for describing format is in large part why the web succeeds.
Notably however this first web doesn’t have many of the features we would consider critical:
It doesn’t have graphics. No cat pictures.
As well, it doesn’t have scripting. No pop-up ads - in fact no advertising at all. No online games, no rich client-side apps. Just words. It’s not exactly the web as we think of it today.
The web at this point is a "thin client model" - with computation performed at the server. Web servers such as httpd are effectively the application, and the client side is merely a presentation layer, to paint the display and collect user input. Of course the web is an “app” but it was only “one app” or “few apps”; a few server side experiences that were carefully curated; effectively store-fronts and proxies and agents for universities or larger businesses.
Because of this in its first incarnation the Internet is bifurcated. There's a World Wide Web for people, and an Internet for machines.
Images
The first person to probably post a cat picture to the web is likely somebody on the team at Mosaic - four years after the first web browser. It’s a momentous occasion that likely passed unwitnessed but still acknowledged.
Scripting
Computer Scientists Alan Kay and David Smith (with Squeak and later with Croquet) were already articulating a vision of personal creativity that wasn't limited to words but to the democratization of programming. In fact Alan positions himself in somewhat opposition to Engelbart’s vision. He wants to empower people to have a "dynamic medium" in which they can personally create any software they can imagine. In his mind if it was a medium "it had to extend into the world of childhood". Everybody should be able to be Engelbart or Tim Berners-Lee. Here we see a hunger to let people speak with computation, to expose real power.
The web picks up scripting in the form of both Java and Javascript in 1995. This signals a transition point for computation as a whole. Macromedia Flash itself also emerged around this time - from work by Charlie Jackson, Jonathan Gray and Michelle Welsh. It’s another subtle yet critical moment for the web.
Contemporary figures in the transition to an interactive web are Marc Andreessen and Brendan Eich. Marc leading Netscape and Brendan writing Javascript for Netscape and much later starting Brave.
This movement towards an interactive web hasn't been without some amusing push back. One notable project was NoScript which promised to disable Javascript in your browser. With more and more of the web being represented by code, and less by declarative text, it seems an utterly hopeless luddite vision of the future.
Folks like Tim O'Reilly noted the transition at the time; inventing the moniker of a "Web 2.0" ecosystem or “Internet Operating System”. Noting improved user agent powers such as better handling for payments, speech recognition, location services, identity services. We also see general purpose computation move to edge devices in general with the rise of the walled gardens such as App Stores and the rise of persistent web apps and even the success of WASM, Kubernetes and Cloudflare Web Workers.
A web that does work rather than describe work
All of these efforts start to change the meaning of the web away from merely displaying text to performing actual work.
It’s not entirely clear if this pressure is conscious or human or if simply part of some innate quality of evolution (and it’s worth someday asking these folks what exactly they imagined would happen when they let anybody write and share their own interactive apps easily).
We’d already seen cloud computing, and an idea of “edge computing” (where computation moves to the edge of the cloud to lower the latency for end user interaction). But we start to see now is computing at the “edge of the edge” - just in time computing delivered to the client device for the lowest possible latency interaction.
And as computation moves to the client, is durable and persistent, this in effect starts to fulfill the promise of a user agent - as in “an intelligent helper”.
This is something we also see with the Apple App Store or the Google Play store; a personal computational platform that can run novel apps and experiences just in time to do novel work. You can go to Mato Tipila and download an app while you’re walking around and get enhanced information about the park; effectively receiving a just in time app delivery with a context aware reasoning about your needs. Or you can download a digital ruler and use it to measure the length of a hallway; again a real time capability that is doing actual work, based on a whim that you had, and were able to satisfy almost just by thinking it.
A fossilized class of applications
The big applications on the web are rapidly becoming obsolete, fossilized in place; they’re left adrift.
Facebook, for ordinary users, only supports legacy media (photos and videos); it doesn’t allow users to easily push their own rich apps out to the edge.
Brewster Kahle runs the Internet Archive "a non-profit library of millions of free books, movies, software, music, websites, and more". But how does one index or search a web that is made up of free floating digital agents that express highly dynamic behaviors? Does Brewster's job become more like Craig Venter's - to merely sample the DNA of this new ocean?
If you rely on assistive text in some way, how do you even consume a web that generates its words dynamically rather than statically?
How do search engines search a web where the content is produced dynamically and contextually? What does search even mean in this case?
We are shaped by our structures
Manuel De Landa writes that "we live in a world populated by structures - a complex mix of geological, biological, social and linguistic constructions that are nothing but accumulations of materials shaped and hardened by history."
We take for granted a digital web context that can be used to read the news, examine depictions of other primates, look up recipes, share sentiments with friends, but as well, deliver computation to your device to reconstruct a model of the room you are in, or play a game, or even become a yoga coach, or advise you on your best navigation route, or become a musical instrument; interpreting your gestures as performance in an aesthetic manner.
But this isn’t the web that Tim Berners Lee envisioned. It is a web that doesn’t merely forward our gossip to each other; that is merely our memories, stories or advice.
It is a web that embodies work, that does work, a distributed web, where computation itself surges to the edges. In a sense our words are magic now, we don’t just have a wise voice whispering in our ears but we have a wise hand helping out. We can conjure our own durable and persistent digital helpers, and have a cloud of assistants that proxy our will; that alert us if our pets stray, or if a rare purchase shows up on ebay, or when it is time to wake up.
The Near Future
Browser companies such as Mozilla phrase their role as providing a “user agent”; or a gateway to the web. A similar emphasis can be seen in mobile desktops as well; with carefully curated apps and a rigorous vetting process. The focus in these cases is on protecting and empowering the user with respect to a virtual cyberspace.
Today our software is largely on demand, we open our phone or device and invoke an app, or go to a web-page, do some work, and then close it down again, but in the near future these apps will be always running and always available.
But future apps will likely run in a software container that is more like a car computing platform. There will almost certainly be some kind of microkernel architecture with many simultaneous processes running at once, each very carefully sandboxed, each collecting and gathering contextual information and communicating with each other and the user.
We will have apps that are providing ambient awareness of services around us in space, that are illuminating things around us of interest, that are acting as memory aids and communication devices; showing us traffic lights, translating languages, showing us the co2 pollution levels above a building. Multiple separate services written by separate companies will compete to paint to the same user display. In fact future browsers may not even include the ability to render HTML.
Like the browser of today however, the future browsers will very much also have an emphasis on cross-platform portability.
It’s important to get this right because the web is more than our shadow or data exhaust. We increasingly live in an architecture of the web. The web is a concrete artifact that increasingly defines us; it increasingly bounds what we are. Certainly our children will think that life before their hyperweb was solitary, poor, nasty, brutish and short. That Kim Kardashian had only 218 million followers may be bewildering to them.
We can observe already in our time that the web we know is fast becoming a sea of computation; digital agents at the edges that talk to us and each other. A web that is itself almost alive.
We can conceive the future web as a service that runs agents, that is more focused on computation than merely text, but what is meant by computation itself? Is there something further beyond what we see now?