stronglink
stronglink copied to clipboard
A searchable, syncable, content-addressable notetaking system
Our search indexer currently uses the Porter stemming algorithm from SQLite FTS3. We've already tweaked it to ignore underscores, but it still has several other limitations, mainly regarding languages aside...
Ironically our blog front-end might be difficult for existing search engines to index because of how it relies on queries for navigation. A motivated search engine might page through all...
We should be able to gracefully handle file deletion. Deletions will never sync between repositories. Deleting a file only deletes it from the local repository. This is critical for backups,...
For images, this requires a fast, portable, and hopefully secure graphics library that we can use to generate thumbnails. For other formats, it gets fairly complicated. These preview generators should...
Right now the upload page "works," but it's ugly (completely unstyled) and difficult to use, because we trust the MIME type the browser tries to guess. It also doesn't convey...
Our Markdown editor is pretty bad. From what I've seen [StackEdit](https://github.com/benweet/stackedit) might be the best option, but I don't know how it handles Common Markdown or how easy it would...
- We should build gzip versions of static files, and the static file server should prefer them if the request accepts gzip encoding - We should consider compressing dynamic responses...
To support files with content embedded from other files, we need to do a few things: - The preview generator needs to resolve the hash URI (which currently it does...
We're using libuv so we get a lot for free. The main problems that I can think of: - We do path concatenation using `/` as a hardcoded separator -...
- Native db_get/put/del support, instead of using the transaction cursor in all cases (get rid of db_ext) - Unify LevelDB and RocksDB wrappers - LevelDB nested transactions (needed for improved...