back to notes

Web Browser History

History is a DAG (like git), with all actions and information required to display the content stored - by default ephemerally, but can be 'remembered', which is permanent. Support branches and tags (bookmarks)?

Storage might be permanent, but the browser will respect cache headers when revisiting url (adding a new node, since graph is acyclic), old remember node is automatically forwarded if there are no significant changes (might have to do a diff of text or image content to be certain).

(Blue sky - p2p archive.org). Key is fast history search, storage metadata, and external program usability (I'm not going to be responsible for that many features!). Best way to do this is following current and existing standards and conventions, for storage, simple and reversible transformations.


last updated march 2013