Thursday, October 4, 2012

universal web comment protocol needed for tunable search

Email and the web triumphed by defining open decentralized protocols that in principle anyone with a computer could use, and still can. Unfortunately, they are a bit high maintenance, so most people do not run their own server with a direct connection to the internet backbone. Most use an ISP. The ISP can spy on them.
Web 2.0 enabled crowdsourcing. We contributed to the construction wikipedia.
Web 2.0 deprecates protocols. Google, facebook, twitter etc. capture our comments and wall off our/their content from the rest. Every web site wants you to get a username and password to make content for them. AOL's walled garden strikes back from the grave! Governments like China love to target the centralized data vaults. Mark Zuckerberg is the devil. He wants to own our data. Faceless... bookless.
We need a truly decentered protocol for commenting and collaborating. I should be able to publish comments on anything, and find useful comments on anything, one system everywhere. Wikipedia should be like a torrent, a truly decentralized peer-to-peer document perpetuated by usage. Similarly with Google's web index. Facebook is just training wheels for the web.
If anyone came up with a better search paradigm than Google's could they resist becoming another Google? Can the web return to its internet roots, or is there no turning back?
Diaspora*, Identi.ca, Friendica, tent, and the open microblogging standard are Daviding against the web Goliaths. I guess there is still some hope.

Why did delicious fail? My comments should be on my server in a peer-to-peer network, and searchable from google or any search engine. I should be able to comment on any URL or URI. Authentication based on public key crypto web of trust.

I should be able to tune my searches on google depending on who I consider an expert. If you are my only expert, stuff you like would top my search results, stuff you ignore would be in the middle, and stuff you dislike would be at the bottom. Or I could mark you as an anti-expert - stuff you dislike goes to the top. Or I have multiple experts, and their attitudes vote on my search results.

Google had the right idea - people's use of the web should self-generate the rankings of pages. But because both humans and bots inhabit the webosphere, this didn't work out. Can we fix this with pubkey crypto/web of trust and tunable search? It doesn't matter if x is human or a bot if I can make x an expert or antiexpert. Otherwise we need a turing test or captchas.

Bots may end up with even better reputations than some humans, by aggregating humans, or by evaluating mechanically. Biz model for bloggers?

Reverse reputation lookup - who rated this page highly? Who uses this page a lot?

No comments: