The culture's (and my own) understanding of large projects that don't follow a benevolent-dictator model is weak. Most such projects fail. A few become spectacularly successful and important (Perl, Apache, KDE). Nobody really understands where the difference lies. There's a vague sense abroad that each such project is sui generis and stands or falls on the group dynamic of its particular members, but is this true or are there replicable strategies that a group can follow?