
I would like to propose an augmented structure for PG. The aim of this structure is to realise Greg's stated desire to include ebooks that are contributed outside of the normal WWer workflow. The basic premise is that the PG library as it stands becomes just one of multiple contributor libraries hosted by PG, with the gutenberg.org interface rejigged to catalogue all of these libraries. In terms of customer interface, the customer gets a searchable catalogue of works. When they access a work, they are presented with a list of which contributor libraries contain expressions of that work, each with some blurb describing the properties of that expression. Thus, for example, the "Huckleberry Finn" page might list two expressions in the WWer library, one in the DP library, one in the Adcock library, an RST derived one in the Marcello library and a ZML derived one in the Bowerbird library. Library owners would be responsible for maintaining their own libraries. Anything in any of the libraries is fair game for other contributors to build on. Contributors may choose to pool resources and maintain common root versions/ master formats, e.g RST advocates might maintain a set of RST files in Mercurial, possibly even grouping together to build a common RST derived library. Customers would be able to list all the books in a particular library, so if they thought the Bowerbird "Huckleberry Finn" was the best formatted ebook they have ever seen they would be encouraged to see what other Bowerbird ebooks there were. If they thought it was awful, they would avoid Bowerbird works in the future. You could even allow customers to blacklist certain libraries from their future results. Library owners would also get the ability to recommend expressions in other libraries to anyone who is visiting their library, creating a useful reputation chain. The WWers should not be upset as nothing is being erased. They have the biggest library, and they are already entirely responsible for its maintenance. It is already maintained to strict standards, and if those standards stack up, it should be the go to library. DP should be pleased. They would get to do CSS3 and make their own epubs if they want to, and they would get to maintain their own library, something that has been talked about wistfully on the DP forums for a _long_ time. Individual contributors should be pleased. They can produce work to specifications driven by their own beliefs in the formats that they believe are superior, and if they are right about the quality of their work being superior, they will garner a following.