E2 Reader- Implementation
The E2 Reader is being implemented as a Web service that is open to invited testers and interested persons in the E2 community.
Phase 1
Take all of the links that are provided in the E2 XML tickers and organize them into hierarchical menus in the sidebar as the interface for exploration. The top-level menu organization is Content and People. It currently looks something like:
- Content
- More by this author
- More of this node type
- Similar topics
- Others recommended by same recommenders
- Flow (softlinks)
- New
- Recommended
- Staff Picks
- Stuff you might like
- Collections
- Reading List
- Power search
- People
- This Author
- Recommenders (who cooled this)
- Followers of this author
- Followed by this author
- Following me
- Followed by me
- Who I might like
- Groups
- Me
The menus under Content and People items provide links to four contexts: the current writeup displayed, the current node, a recommendation space, and the reader (you). Think of each context as a cluster of immediate destinations that you can reach by a click from where you are at the moment.
The bolded items above can be implemented with the current XML interface that E2 provides for client programs. I have them working now. The others are new features that need information the XML interface does not provide. Phase 2 is about implementing those features.
Phase 2
The purpose of the second phase is to test ideas for discovery and recommendation that are not directly provided by the current E2 interface. The information on content and people for those links will have to be scraped from ordinary E2 pages or generated by new discovery and recommendation mechanisms.
Obtaining the needed information efficiently is the first issue; the second issue is storing the information for efficient retrieval. The data could be stored within E2, but it would require building new database indexes or new database tables, which I think is not going to happen.
The best immediate solution to both problems is to store the information outside of E2 in a different sort of database structure that is called a graph database. The graph database is particularly good for this purpose because its structure reflects the natural linked structure of the data and the structure of E2 Reader navigation, which are the same.
Previous: Introduction | Next: First revision