This is from two emails I sent yesterday. I'm not ready to post something complete about it yet, because my mind isn't made up. Still, worth sharing::::

I’ve recently become convinced that we — the world — made a huge mistake by starting with the Web as the first widespread application on the Internet. Specifically, we made a bad decision by basing the Web on client-server rather than peer-to-peer. Client-server is a mainframe legacy. It’s a slave-master arrangement.

Peer-to-peer is what the Net was designed to be, and still is, at the protocol level where it came into existence and still persists. We need to re-frame our thinking there, no matter what we build.

…and…

...take a look at Brad Burnham’s post on Union Square Ventures’ investment in Protocol Labs. Note that he's wanting to invest in protocols rather than platforms.

Also take a look at what I said in Giving silos their due, and to the comments as well, especially Michael Elling's. Phil Windley’s Decentralization is hard—maybe too hard is also a good response.

My point with the last two paragraphs is that my mind is not made up on whether or not a platform approach is a Good Thing, though I am beginning to think there are Better Things.

I am convinced, however, that we went way wrong with the Web as a platform by basing it from the start on client-server (aka calf-cow), which (to leverage McLuhan) retrieved the mainframe, which retrieved the monopoly, which retrieved the feudal system.

Hence the headline.

I welcome your thoughts.

[Later...] Best so far is from Dave, via this tweet back toward Fractional Horsepower HTTP Servers, posted twenty years ago. In a second tweet he adds, correctly, A kitchen is client server too, but we scaled them so bigcos don't own them all. It's more that we're lazy. We like TV. McDonald's.

I also just added an image, of silos.