Sorry for this being a separate email. I normally follow this list using the digest format, which does not suit itself to replies. While I have not been active in many years, I am still around.
My opinion is that the "benevolent dictator" model is the best for a project like gramps. A project needs a clear vision and that is the responsibility of the BD.
As examples, both the Linux kernel and Python follow the BDFL ("benevolent dictator for life") model. Both of these projects are highly successful and have a clear direction. Many people contribute to the projects, but in the end, the BD keeps the project focused.
A rule by committee model tends to pull the project in many different directions, and the project loses consistency. While each individual decision may seem to be a good point decision, no one is there to keep consistency and provide a long term direction. This model also allows the project to be hijacked and taken into tangential directions by a few. In the early days of the project, we had many who insisted that we had to switch to the QT or TK toolkits, or adopt the Gedcom XML as the internal format, or abandon the desktop model and focus solely on web based projects.
The benevolent dictator model means that if you want a significant change, you need to be able to convince the person at the top that your idea is worthwhile. Chances are, if you can't convince the BD, the general users won't be happy either.
When I abdicated the role of BD, I could have set up a committee to replace me. However, I chose instead to have Brian Matherly replace me. He had a good idea of the direction of the project, he had excellent leadership skills, and he had good people who could advise him. The model has continued to work well, considering the project is in good shape, and has been highly active for almost 17 years.