Dan Connolly's tinkering lab notebook

On OpenID and comment policies

Writing to this journal involves Yet Another Password. Sigh. Oh for OpenID support in drupal. Some folks seem to be working on it.

Of course, if our policy aware web project gets anywhere near where we hope to go, it should provide enforcement for very flexible policies.

What's a good comment policy?

  • You can comment if your homepage has pagerank>K
  • You can comment if any of my friends say you can comment

Hmm... I'd like something a bit more scalable. I wonder if the advogato trust metric can be distributed. I seem to remember some distributed time protocols that were robust up to massive collusion, but I looked hard for them in citeseer and couldn't find them.

postscript: Karl notes Russell Beattie on Anonymity. I'm not sure I agree that most people are anonymous by default. I think that by default, people identify themselves whent they speak, but they don't speak to very many people. Until the last few generations, most people lived out their whole lives in the same city and never spoke to anyone outside that city.