Steve's excellent analysis of how the Network Effect worked against Mojo indicates a social-fix for Mojo++, ie: make it easy to get on and get content. Ie, get them hooked. They'll at least be autosharing stuff they've downloaded. After they're hooked, folks may feel like contributing their time/attention to injecting new material, and to get "unlimited" access folks will have to contribute. But you have to get them hooked first.
I daresay an empirical study of the successful P2P systems' user-behavior would show this. Pure technologists may abhor the social/human factors issues that arise (eg in crypto UI, P2P config) but we should regard them as problems to be analyzed and solved. At 09:32 PM 11/23/02 -0800, Steve Schear wrote: >I was part of the team and I respectively disagree. Sorry to sound a bit >like Burke's Connections series, but Mojo failed because of network >effects. To wit: >1 - key features needed to automate the creation of descriptive header >meta-data were not delivered, which caused >2 - content to be too sparse, which caused >3 - too few new users to take an active interest in contributing new >content (since most wanted to download first and only later might have >contributed new content), which caused >4 - the number of Mojo brokers holding content to be too small to give >adequate persistence to content already contributed and increased churn , >which caused >5 - See item 2
