urlspace is a way of organizing a bunch of documents that could become a static/dynamic content management system, provided I write the actual code.
Documents are nodes in a tree. Nodes have attributes.
The tree can have multiple representations, some isomorphic (e.g. files on disk ↔ in-memory representation ↔ database), some not (e.g. in-memory representation → HTML files).
The tree is mapped to an URL space (see On short URLs for some related thoughts). This mapping should probably be one-to-one; many-to-one is more flexible, but requires extra choices (e.g how do you represent multiple nodes with the same URL?).
Goals:
- finish something; files on disk → HTML files at a minimum
- write as little code as possible (libraries are OK)
In scope:
- node types
- node ordering
- inline references to other notes
- (later) resources (node attributes that are not stored with the tree)
- (later) error pages
- (later) live preview
- (later) Atom feeds
- (later) sitemap
- (later) image handling
- (later) full-text search
Out of scope:
- versioning (it is representation specific, e.g. git for files on disk)
- web UI (files on disk should be enough)
- internationalization