Rebooting Web Comments

(A mathematical comment on Tim Gowers' blog)

An acquaintance told me yesterday that she's quit reading a community website that she used to enjoy. Why?

I can't stand the commenters. I can't believe the amount of time on people's hands, and how nasty and petty they are. I think people on public forums should be required to use their real names when commenting. If they did I think there would actually be some quality of discussion, and it might be useful.

Anonymity or pseudonymity can indeed bring out the worst in people. That's true in both physical and virtual realms. I've long believed that with rare exceptions (for whistleblowers, political dissidents, victims of abuse) we are best served by online discourse among parties who use verifiable names. I wish strong cryptographic verification were feasible. For a host of reasons it isn't yet. But sites can at least use the OAuth mechanism to leverage existing Facebook or Twitter or Google or Microsoft identities. OAuth is easier to implement than it was a few years ago, and it's way better than using self-assigned screen names. You'll think twice about that nasty or petty comment if it's backed by an online identity that you've made a long-term investment in.

More durable identities will help improve web comments, but comment systems are broken in an even more fundamental way. You don't own or control what you say in comments that you post to forums, blogs, and Facebook. Your words are scattered all around the web. Things you can't do with that dispersed corpus include:

- Search it effectively

- Assure its long-term preservation

- Monitor its usage

Imagine a different way. You visit a site, read something, click the comment button, and launch an edit window. Nothing new there. But here's the twist. That edit window is wired to your personal cloud. That's where your words land. Then you syndicate your words back to the site you're posting to. What appears on that blog, or in a forum, or (in my dreams) on Facebook, is a view of one piece of your permanent digital archive.

There are technical challenges of course. This scheme will require a standard protocol enabling sites and personal clouds to collaborate in this way. And plenty of caching so that the site doesn't bog down trying to display a page that pulls in comments from many personal clouds. And given that caching, a backchannel that conveys usage data from the site to those clouds.

We can do this stuff. And there's money to be made doing it. Today, for example, I pay almost nothing for my WordPress blog -- just a token annual amount for domain name redirection. I'd pay more than that to connect my personal cloud to the entire WordPress ecosystem, and still more to extend that connection to the wider world of sites where my comments can appear.

Some of the best customers of such a service will be academics. There's a whole generation of frustrated young academics who want to regard blogging as a professional activity but can't. Why not? With rare exceptions, blogging doesn't count for reputation or tenure.

My favorite exception is the Polymath project, which began when Tim Gowers, an elite mathematician, asked his peers: Is massively collaborative mathematics possible? The answer, which Michael Nielsen explores in his book Reinventing Discovery, was resoundingly yes. Gowers, along with fellow Fields Medalist Terry Tao and others, have since used their LaTeX-aware blogs to host large discussions that have cracked some hard problems.

When you're a Fields Medalist you can get away with that. But most academics aren't and can't, so they blog in their spare time. That's a shame because what the blogsophere so effectively enables -- open and free-flowing discourse -- embodies the academy's highest vision of itself.

In a meritocracy based on publication and citation, of course, you have to be able to publish in a controlled way, and you have to be able to count citations. Today those functions are provided centrally, by universities and by publishers, in ways that drastically limit open and free-flowing discourse.

We need to transform those central repositories into hubs that syndicate to and from personal clouds. It will do wonders not only for academic discourse but also for the entire web.