Person/codekiln/GitHub/logseq-gardener/Project/Goals

• logseq-gardener Project Goals

  • garden aims to create the fastest incremental and lazily-cached back-end for three primary use-cases:
    • 1 - an garden CLI for humans, agents and scripts to query and update Logseq/OG-style Knowledge Gardens
    • 2 - an garden lsp CLI-powered, back-end for an LSP that can enable LazyVim to provide a Logseq-style user interface
      • an editor plugin starts garden lsp and keep that process around; no TCP sockets or ports or server set-up.
    • 3 - garden publish a cache-backed static HTML per-page output for SEO and indexing
  • garden-core should provide a parser, a graph model, and caching affordances.
    • It's meant to be as performant as possible, respecting that caching can start in the background the moment the engine knows about a garden, then incrementally enriched as each command is executed.

      • graph activation
              ↓
        background mldoc outline parsing
              ↓
        in-memory page/ref index
              ↓
        [[ autocomplete

  • The implementation should provide support for concurrent writes to the same file and even the same block, using CRDTs as a front-end to the cached backend.