dave peck :: software, music, nerdery

On the Cusp: Decentralization

One of my sporadic posts where I guess about near-future technology trends.

The term decentralized is intended to stand in contrast to the centralized nature of modern social networks and SaaS solutions.

The centralized model has led to immensely valuable services. Twitter, Facebook, and many enterprise-facing services are highly evolved organisms, well adapted to the technological ecosystem in which they thrive.

Yet, modern tools like the block chain, BitTorrent, DHTs, PubSubHubbub, and even oldies like RSS, present us with the opportunity to build curious new creatures with characteristics wildly different than the services we’ve seen before.

I’m explicitly not interested in repeating the past. Decentralized systems meant to mimic centralized systems (for example, Diaspora) ignore the tradeoffs and opportunities inherent in our modern tools; they will fail as a result.

Likewise, I’m not interested in replacing centralized services. Decentralized systems will always have unique strengths and weaknesses and will always satisfy radically different user needs than their centralized counterparts. The first decentralized service to gain real traction will succeed precisely because it embraces its technological nature.

As an aside, I wish there were a better word for this trend. Most client/server architectures built on the TCP/IP stack are at some fundamental level already decentralized. The web is decentralized; Facebook, a specific site on the web, is not. This said, P2P is not nearly a broad enough term, and distributed has specific engineering connotations (think: CAP theorem) that don’t necessarily apply. No wonder the community has decided to adopt decentralize for now.


I don’t use Facebook much. I have an account so that I can write software that integrates with it; as a result, once in a blue moon I find myself reading friends’ posts, or writing a quick one myself.

My Facebook profile consists mostly of photos I’ve been tagged in by my other, more Facephilic, friends. It is fragmented and vague, like a funhouse mirror seen through a fog. My Facebook profile doesn’t look like my own life to me.

And yet, there’s tacit assumption that Facebook is where our digital lives are. I was sharply reminded of this when Amy and I joined a support group for new parents here in Seattle. Facebook is how everyone communicates and, alarmingly, how everyone forms an impression of others. What impression might our new parenting friends have of Amy and me?

This is yet another reason why I increasingly want to own my online presence.

Twitter Mainland

Brent Simmons deleted all his old tweets. He wants to own and control his writing, including his microposts.

Manton Reece chimed in with an interesting reply:

Justin Williams joked that Brent and I are now the sole inhabitants of “Manton Island”. That’s funny but it’s actually backwards; it’s Twitter that is the island. Everyone is there, though, in an overpopulated mess, so they don’t realize they’re cut off from the rest of the world — the open web, designed 25 years ago as an interconnected system of countless islands.

I think it’s a noble goal to build systems that let people express themselves while retaining full ownership and control of their expression. I’ve spent a surprising portion of my recent downtime thinking about such systems, and I’m slowly consolidating my own output under one roof.

But there’s an uncomfortable truth to Justin’s joke: most people don’t care about ownership and control. At least, not principally. To them, Twitter isn’t an overpopulated island: it’s the mainland. Everywhere else is the vast, quiet sea.

If we want people to set sail for brighter shores, we’ll need to do more than replicate the features of central silos in distributed fashion. As table stakes, we’ll need to provide world-class user experiences for reading and writing, and we’ll need to have a story for discovery so the network can grow.

But most importantly, we’ll need to offer something new to do on our islands that can’t be done, or done well, anywhere else. Network effects are the lifeblood of any social network. Our travel brochures will need to offer compelling new reasons to visit.

I suspect this is why the systems I’ve played with so far have failed to thrive. Some slavishly copied their progenitors, providing little added value. Some offered bridges to ease transition, only to become mere appendages to their centralized counterparts.

This is our challenge and our opportunity. Twitter is a highly evolved organism, well adapted to the centralized landscape. Distributed systems are fundamentally different terrain with fundamentally different character. What new and compelling organisms might thrive here?