Client Server Aggregator
I personally prefer the 3-paned reader over the browser aggregators, but I'm tired of being chained to a single PC in order to efficiently consume my feeds.
I suppose that one could make AmphetaDesk or RadioUserLand-like aggregators operate in a network transparent remoted environment, but as mentioned previously, I prefer the 3-paned reader approach and none that I have found supports network transparent separation of the user interface and back-end repository.
Is this an application implementation detail or is there a case for specifying aggregator functionality as a set of client and server APIs?
The functional requirements are:
The back-end system shall act like a central, IMAP-like, repository (for feeds, not e-mail)
The back-end system shall store subscriptions
The back-end system shall store the read/unread state for each item (based on GUID or some such)
The back-end system may store the actual feed content (minimize consumption of bandwidth from the feed source)
Separating the user interface and back-end repository offers the most flexibility and would work in the case where a user wants to run the client and server on the same machine (after all this is in essence what the browser-based AmphetaDesk and RadioUserLand already implement).
Comments and suggestions
Comments, suggestions, and errata are welcomed. -- [CaS]