Serves documents from a particular folder. It is intended to work together with a regular HTTP server folder.
A good way to do it would probably be to start with an algorithm that extracts a spanning tree. This tree would then be traversed, and the links changed from html://host/document to thout://hash_function/hash links. The tree might need to be saved for several reasons. (E.g: repeatability, optimisation, incremental updates, etc...)
In case the folder in question contains several disjoint sites (i.e, the digraph generated by considering the pages as nodes and the links as arrows is disconnected), the algorithm would need to generate a spanning forest. Thus, it needs a subroutine to insure that the contents of the folder have all been traversed after indexing each connected component.
There is more information on the roadmap, including a worked out example.
Serves documents from a particular folder. It is intended to work together with a regular HTTP server folder.
A good way to do it would probably be to start with an algorithm that extracts a spanning tree. This tree would then be traversed, and the links changed from html://host/document to thout://hash_function/hash links. The tree might need to be saved for several reasons. (E.g: repeatability, optimisation, incremental updates, etc...)
In case the folder in question contains several disjoint sites (i.e, the digraph generated by considering the pages as nodes and the links as arrows is disconnected), the algorithm would need to generate a spanning forest. Thus, it needs a subroutine to insure that the contents of the folder have all been traversed after indexing each connected component.
There is more information on the roadmap, including a worked out example.