I have recently been reading Ralf Sudelbucher’s blog about Pile, a concept of storing the relationship between data insted of the data itself, the data can always be reconstructed from the relationships.
Anyway after having a play with his version of the pile engine, and agent I have decided to have a go myself. To be honest it started after I had copied in Ralfs code, following his design and writing a few unit tests to ensure I understood what he was doing. and found that the engine worked, but I could not find my data, I can reconstruct the string when giving the agent the document handle but I could not find any test using the search.
So after looking I have decided to have a go myself. as things progress here I will post bits and pieces and we shall see how we go, hopefully by the end we should have a working engine, wihich can use disk space insted of just memory.
So the plan is to construct a working pile engine, and text agent (think I will design the engine to store text, this may help in storage etc.)
fully functional search
OK that will do for the minuite.
The Butler Did It 