Sometimes it’s worth reading the comments. For instance, take this comment on the web’s consolidation from Mike Caulfield.

You look in 1993 and see Guido Van Rossum and Berners-Lee arguing that instead of an IMG tag there should be a general “include”, that would allow you to pull together pieces of multiple sites together from multiple MIME types. Twenty years later, there’s still no include.You see Shirky and Weinberger talking in 2003 about how the web was designed to connect pages, not people, and the groups forming were essentially hacks on top of that. But that power to connect people doesn’t get built into the protocols, or the browser, or HTML. It gets built on servers.It’s almost like the web’s inability to connect people, places, and things was the ultimate carve-out for corporations. [I]f the connections have to live on a single server (or server cluster) then the company who controls that server wins.

The lack of an <include> tag led to Pinterest. No method to connect people created Facebook. RSS’s confusing interfaces contributed to Twitter’s success. Any guargantuan web company’s core value is a response to limitations of the protocol (connection), markup spec (description), or browsers (interface). Without proper connective tissue, consolidation becomes necessary to address these unmet needs. That, of course, leads to too much power in too few places. The door opens to potential exploitation, invasive surveillance, and a fragility that undermines the entire ethos of the internet.

[Edit: APIs were at first a patchwork to resolve the shortcomings of protocols. They let data flow from place to place, but ultimately APIs are an allowed opening to a private dataset—a privatized protocol. The halcyon days of Web 2.0 were a short lived window of benevolence that eventually closed.]

If a fifth of the planet signs on to Facebook each month, why shouldn’t a neutral version of it’s functionality be built into the protocol, markup spec, and browsers that drive the distributed web? (An argument could be made that Facebook is already trying to do the inverse—turning the internet into Facebook—with its internet.org campaign.) We should view the size and success of these companies as clear calls to recreate their products’ core functionality and weave it into the fabric of the web.

What if tech companies were field research for the protocol? This may be a dream, but it’s our only hope to refragment the web.