Older blog entries for fen (starting at number 2)

Pymmetry and Bram's "trust" code have gotten me (finally) to spending a little time with Python. It's fun and easy though I still have some of the steep part of the learning curve to go up. Emacs integration seems good, but I can't seem to find the key-binding to evaluate e.g. a test def in the file I'm working on. I'm sure there's a way...

Working with existing trust frameworks has got me thinking about how cool the OpenPrivacy reputation management framework is. It's designed so that trust metrics - such as Pymmetry or Slashdot's moderation - can be plugged in and evaluated *themselves* on their reputation. So a community that uses e.g. Pymmetry today can easily switch, if and when a better trust metric (or a newer version of Pymmetry ;-) comes along. All pre-existing identities, certification, and reputations would remain intact, perhaps translated (at owner discretion) to the new system.

Think of it like a currency exchange, but with reputations. No one can force you to start using a new currency but if all your friends move to France, you'll want to start using francs. The Reputation Management Framework provides a plug-in architecture for Reputation Calculation Engines that make this sort of "reputation-exchange" feasible. And since reputation-exchanges are not zero-sum, you actually get to keep your old reputation, too!

We're putting the finishing touches on the documentation, but the code is available now. We're also working on a example system called Reptile (Reputation-enhanced portal using Mozilla technology) - check it out!

Fought with Debian today. I had moved up to "unstable" on my personal machine to get access to some new stuff. Guess I've been lucky - and source control is getting better - as I've had no problems. But I want to get it back to "testing" level, which is anything but straightforward.

I've been thinking a lot about 'trust' and 'reputation' recently. (I presented a paper at the Boston Computers, Freedom and Privacy conference earlier this month entitled OpenPrivacy - Enhancing the Internet with Reputations.)

Trust is key to any anonymous network, indeed to any society. Pseudonyms that over time prove to be trustworthy develop reputations that smooth the process of finding the people/entities that are most worth dealing with - within any given domain.

Trust is not generally transitive, as I may trust an auto mechanic with my car but perhaps not to invest my money. Ultimately, we find that trust develops through prior experience and knowledge, is spread by word-of-mouth, is dynamic, and non-monotonic.

There's precious little useful research in the computer science literature on trust and reputations, perhaps because there's a need to understand some background in the social sciences as to what it is and how it works. But I'm excited that more effort is going in this direction. Here's a short Bibliography that I have compiled. Suggestions are welcome!

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!