SpadFS is an attempt to combine features of advanced filesystems (crash recovery, fast directories, etc.) and good performance without increasing code complexity too much. It uses crash counts instead of journaling (because journaling is too complex and bug-prone), and uses hash instead of btrees for directory organization.
Why I wrote Links
There have been questions why to write Links instead of improving Lynx and W3M. Links and Lynx have absolutely different programming style. Links uses callback mode to manage most actions - when we want to (for example) make connection, we call function make_connection and pass pointer to functions that will be called when the connection is ready. make_connection registers request and returns immediatelly. (Netscape does it the same way). In contrast Lynx and W3M use blocking calls for many actions - they call make_connection, it waits until connection is done and then returns. As a result, Links easily manages more simultaneous connections. Lynx and W3M don't and _never_ will.
Lynx will never have features like continuing loading page even if user selected 'Back', downloading files in background, preloading documents, revalidating cache in background. Links will.
Joining Links and Lynx is hard. They can reuse html parser and table code in Lynx, but the core is completely different.