I recently stumbled on a piece by Dieter Bohn on The Verge titled “And now, a brief definition of the web” that got me thinking.
He starts with the question: “what exactly is the web?”
Traditionally, we think of the web as a combination of a set of specific technologies paired with some core philosophical principles. The problem — the reason this question even matters — is that there are a lot of potential replacements for the parts of the web that fix what's broken with technology, while undermining the principles that ought to go with it.
I like this take. “The web” isn’t solely a stack of technologies (URLs, HTML, CSS, JS). It’s also a set of principles—principles, as I wrote, imbued into the web with its birth:
the ethos present in the person who birthed the web was instilled into online culture from the very beginning. Berners-Lee believed the web should be open and he gave away the protocols that powered it. There was no patent. No licensing around who can and can’t create a website. It was all put into the public domain.
This principle of open access to information for everyone is part of what we mean when we say something is “of the web”.
Bohn weaves that idea into his proposed working definition for determining whether something is part of the web.
To count as being part of the web, your app or page must:
Be linkable, and
Allow any client to access it.
First, it’s interesting how much of what’s in modern app stores fails the first test. The link is powerful—so powerful that large organizations like Apple go to battle to exert control over what can and can’t be linkable.
But, as Bohn says, “links aren't the complicated part; it's the part where your thing should allow any client to access it” that’s hard.
You can run through all the web-like things in that list above, look at that two-part test, and just say straight up that these things don't count as part of the open web.
Android Instant Apps: only work on Android. Not the web. Facebook Instant Articles and Apple News: pay no attention to their weird URL redirecting and HTML-esque code, they only work on their respective platforms. Not the web.
It’s funny: if you apply this test to a few things traditionally considered “of the web”, they might fail.
Or, put aside the question of access for an “outdated” client. What if you have a “modern” client and you visit a web page (like YouTube TV at one point) that says “this page only works in Chrome”. It’s a URL you type into a browser, so it sounds like the web, but it’s only accessible by a specific client. Is that “of the web” or is it more akin to something from a proprietary app store?
I like Bohn’s working definition, however vague it might be in the details. “Be linkable and accessible to any client” is a provocative test for whether something is “of the web”.