in server route, publication ID is base64-encoded filesystem path (or alternatively URL to public HTTP publication)

[danielweck]

It would be great if this could be arbitrary UUID or database URL scheme (etc.) with backend capable of supported various storage retrieval mechanisms. This is somewhat related to the notion of "fetcher" in the Readium2 architecture (e.g. exploded EPUBs vs. packed/zipped publications), but there are deeper ramifications due to how the decoded base64 string is used to map other behaviours. To be investigated further...