,
14 tweets,
3 min read
Read on Twitter
One of the reasons a #webOfTrust based #pseudonymousReputationSystem has never gained significant traction is that no one has ever released one with enough features to be functional. No one has ever built and released one with enough complexity. A short thread.
I’ll draw an analogy between making the first successful WoT-based PRS and making the first successful combustion-engine based automobile. There’s a certain minimum number of features and level of complexity that you’d need to achieve for your auto to be useful in the real world.
Suppose you did everything right except you left out the auto’s transmission system. Maybe the concept just never occurred to you. Or maybe there was a rush to release a product. Would it work? Nope. It would be a completely useless waste of time.
Or suppose you left out the steering wheel. Or you used unrefined petroleum. Or couldn’t figure out how to get your spark plugs to work. Or how to keep the piston from locking up. These are all details you have to get right for the auto to have any utility whatsoever.
One of the impediments to building a WoT-based PRS of sufficient complexity is that some think complexity means bad UI. The answer is that it doesn’t have to. Put the complexity under the hood where users can ignore it if they desire, or access and tinker if they desire.
I envision WoT based PRS as an open source system with no one in charge. Most users will be casual users: they will never (or rarely) look under the hood. A small fraction (10%? 1%? less?) will be “power users” who will have fun tinkering under the hood making small improvements.
An even smaller fraction will be fully-fledged devs who make contributions to open source code. Power users will be vitally important: they will come up with things like new ratings metrics (eg flag content with some newly conceived label) but will do so via UI’s made by devs.
Another impediment to the initial introduction of a working WoT-based PRS is lack of vocabulary. People can easily talk past each other. Imagine one engineer brainstorming about how to make an engine’s piston not lock up; another says pistons are the dumbest idea ever.
The problem is they each use the word “engine” but the first is thinking internal combustion engine whereas the second is thinking jet engine. If these terms have never been invented, it becomes difficult to discuss the relevant concepts to build something useful together.
Words like reputation, trust, identity are like “engine” in my analogy. People can use these words without realizing someone else may have something very different in mind. To communicate effectively we need to recognize when this may be happening ...
... rather than calling someone an idiot for trying to design a piston into a jet engine’s combustion chamber.
I’m making the case that building a useful WoT-based PRS is largely an engineering problem. But as @iang_fc points out it’s also a sociological one, meaning (I think - Ian correct me if I’m wrong): we need to understand the needs of the user. This is not a trivial task.
For autos it’s not so complicated to understand customer needs: I know even if I engineer a great auto it will be useless to someone who lives in a swamp. But for identity, trust, and reputation it’s much more difficult to understand (or discuss) what these needs are.
(And I haven’t even touched upon them in this thread.) So engineering and sociology must go hand in hand. The what (or the how) and the why. In both cases, nose to the grindstone. End of thread.
