My Eulogy for the Open Web and Old Google
Why did the web grow the way it did?
Let's take a trip back in time: It’s the late 80s and early 90s. By 1993, Windows enjoyed a desktop operating system market share north of 87%, with Apple struggle-bussing along with the remainder. Meanwhile, Server installations were Unix, NT, Mainframes/MInis and the one weird apple Quadra languishing in the corner of the datacenter, dusty and unused. The ‘workstation’ market, largely consumed by Sun by the end, was a mere fraction of a percent of the market for desktop machines.
The world was discovering online spaces like CompuServe, AOL and Prodigy. Sometimes people would be able to see the larger ‘internet’ through the first pinhole gateways between these online services.
Gamers and nerds like me were playing Doom against each other or telnetting to MUDs. There were even some MUDs on Compuserve.1
Anyhow, at the time if you were a programmer or a company making software, once your software passed certain thresholds of popularity, you would be presented with a choice; Sell to Microsoft or compete with Microsoft doing a version of your software, sometimes putting you out of business.
This played out over and over again. MS would buy a company, sometimes that software would flourish under MS like Powerpoint. Other times Microsoft would fail to acquire the company and would compete aggressively with their own version of the product. They would use their desktop power against their competitors. The old joke “DOS ain’t done til Lotus won’t run” was a very real sword hanging over competitors. After failing to acquire Intuit2 they would start their own ‘Microsoft Money’ product, which didn’t really make it.
This was an environment that made succeeding on these operating systems difficult if not improbable for anything but games and the odd productivity/industrial/enterprise app that was often specialized in the extreme. The talk in IT departments was about client server applications and middleware that would work with enterprise deployment systems to allow for greater productivity (etc etc). 3
The first toddler steps of the web and http in the early 90s came as a breath of fresh air to programmers of my vintage. We no longer needed to ship on Windows or Mac. We didn’t have to deal with Microsoft C and the Windows SDK4. We didn’t have to deal with weird Pascal variants to ship on Apple.
We could, using the larval forms of Apache and/or scripting languages we loved like Perl to present fun web pages to the users. We didn’t need to ask Redmond or Cupertino for permission to exist. No one was forcing any of us to use their server, their client, or their operating system.
If you were running BSD (and later Linux) you didn’t even have to subscribe to their ideas about how to run services, all you had to do was output HTTP and HTML over socket 80 and the users could, and would, be able to use your site.
Now, there were commercial offerings for web servers. In the time it would take for Microsoft or Sun to figure out the programming and pricing, then call you back to talk with you about client access licensing and how that might fit within your Select/Enterprise agreement, you’d have as likely as not died of a wasting disease or old age, or questioned what an open ended select client access license even, like, was.5
While the big cos were figuring out the web, the Apache web server would emerge and evolve and extend with the modules systems. It was open source6 software, and no one had their hand out and you could just start serving users. It was a remarkable time. The bridle was off! The internet was how you’d ship to your users either on your internal network, or to users across the world.
For all intents and purposes, the stranglehold that the desktop OS companies had over independent software developers was broken.
It wasn’t perfect, of course. People would lament the lack of an ‘agreed upon ontology’ for web pages, or the ‘stale link’ problem where web pages would go away. Most proposals to ‘fix’ these problems with the early internet looked to set up central authorities that would gatekeep the web. The web was the wild west, they’d say, and many wanted to be the sheriff or governor of this new frontier.
For a long time, about 10 or 12 years, I had the distinct pleasure of having Ian Hickson7 working for me. He had started at google in 2005 after he had spent some time at Opera.
He was working on the next versions of HTML at the time and Google wanted someone who wanted to move the web forward so it could be more competitive with desktop experiences. Back then, Google needed the open web to flourish , as that was what Google crawled, indexed and offered ads next to. We had a ton of efforts designed to make the web better, more responsive, safer and more effective for our users. If the web grew, the thinking was, Google would too.
Google would measure user latency as we knew it had a direct relationship with perceived quality of the sites users would visit. A colleague of mine, Steve Souders would put in an incredible effort at Google and with the outside world to improve the baseline of performance on the web. He would release software to ‘make the web faster’ and that was really the goal, keep the web competitive with desktop experiences.
It was what drove our contributions to firefox (pre-chrome!) and directly led to the creation of the chromium and chrome teams. We were very opinionated about the direction the web needed to go to keep Google competitive. We needed paths to our users that didn’t go through our competitors. It drove the strategy behind Android, Chromium and so many other efforts at the company.
Anyhow, we hired Hixie to work on HTML, and he found his way to my group as I was the only person working for Bill Coughran who had any experience with managing spec writing with folks like the W3C.
Here’s the thing about Ian, he was an absolute machine. In 2004, the w3c was pushing hard on standards for html that were not being , lets say, embraced by Mozilla, Apple or Opera. Ian’s response was a very open source one, he forked HTML and looked at where browser developers were going and created his standard accordingly. He wouldn’t embrace anything that the browser developers were not interested in or actively implementing. Then he and Håkon Wium Lie created the ACID tests of browser functionality.
In rapid order Ian’s combination of *astonishing* productivity on the WHATWG specification and the ACID tests changed everything for the web. Browser creators would ‘compete’ to be more compatible, and it kept the web moving forward for its users and developers. The w3c would , after a bit of discomfort with Ian basically forking their most important work, come to adopt the WhatWG as the next and subsequent versions of HTML. It took a while, but eventually everyone came to some level of peace with it.
About 10 years into Ian working for me, he came to me as he was banding together with some other folks from Chrome to work on something new: Fuchsia (now Flutter).
He had already started working on the docs and specs behind the technology and wanted to move full time to that effort. I remember asking him what would happen to HTML if he wasn’t in the catbird seat? What would happen to the evolution of the open web? He had two answers. The first was that there was a very capable inheritor able to take up the reins from Ian. The second was “You know Apps are killing the web. Fuchsia is our chance to be part of that.”
The iPhone and Android app stores were walled gardens, the social networks were walled gardens, they all existed to gatekeep to their users, you see, and Ian saw before many of us, that this was the case. Instead of pushing the HTML boulder up the hill, Ian wanted to have a go at rocketing past it with open frameworks like Flutter.
As we entered the present era: the open web is getting less open by the day. Countries blocking sites, traffic and applications have become the norm. The splinternet is very real. It used to be that you could count on your internet packets ‘routing around censorship’8
I would assert that for Chinese, Russian, Iranian, European and, yes, American , internet users this simply isn’t the case anymore. And along with these new restrictions on information, the old gatekeepers have risen up to take their … smidge of control….their pennies (or 1/3rds9) of tolls before you could reach your users. It’s decidedly returning to the late 80s, in my mind.
To be honest, I don’t know the answer, the solution, to this problem. I hope the web isn’t lost and we’re not reverting to thin clients in front of mainframes giving us the allowed sites, information and applications that those in power think we should consume.
I am not so naive to think there is much in the way of hope for ‘less’ regulation of the internet and sadder still, I don’t picture Google fighting for the open web at all.
They’ve lost the thread for it, instead they are becoming rapidly more comfortable being the gatekeeper, the toll booth worker, and not the company that I joined that saw an open , vital web as the one that a company like Google could flourish in. I would assert that lacking the passion for that fight, they will surely, and pretty simply, lose it and become another fading former star of computer science chasing marginal enterprise revenue.
Is it too late for Google? I really hope not. I truly hope my former employer can find themselves, that they can fight for the open web and crawl the walled gardens as if their future depends on it. Because, well, it does.
Forgive my musings. Clearly this is an ‘old man recalls odd facts about his early computing life’ post.
A deal that was nixed by the DOJ
I admit this always came across to me as trying to recapture the thin terminal market and wrest control back from Microsoft.
Literally I had a shelf that bowed under the weight of the accompanying documentation for C 6.1 and the Windows SDK. I had to get a thicker shelf until I traded all those docs and discs for an AT&T 3b1, but that’s a different story…
The term hadn’t been decided on yet, but the Apache license was the model for so much of that movement.
And thanks to Ian for checking that I wasn’t misquoting you or wildly off base.
As John Gilmore would say.
It’s seriously appalling and a much longer article…..