This roadmap describes how CrossPoint is moving through the tighter scope defined in SCOPE.md. It is intentionally phased: Phase 0 closed out the commitments already in flight before locking down to the stricter "fill gaps the stock firmware leaves" delineator.
Phases are sequential. We do not start the next phase until the prior one is wrapped or explicitly carried over.
Goal: Land the work that was already in motion under the prior, broader scope so contributors are not left hanging, and so we enter the stricter phases with a clean slate.
Landed in Phase 0:
- RTL support PRs. The in-flight right-to-left work was reviewed, iterated, and merged.
- Dictionary PR. The offline dictionary lookup work was reviewed and merged.
- Bookmarks feature. First-class navigation markers in EPUBs.
Transparent sleep screens.Shelved; not picked back up under the stricter phases.
Phase 0 is closed. The tighter scope in SCOPE.md is now fully enforced. "But it was on the old roadmap" is not a valid argument for accepting a PR.
Goal: Reduce memory and flash usage, clean up the codebase, and land the SDK / HAL generalization work so CrossPoint runs cleanly on ESP32-based e-reader hardware beyond Xteink (X3 / X4), including ESP32-S3 class devices.
Focus areas:
- DRAM and heap fragmentation reduction across the reader core.
- Flash footprint reduction (dead code, redundant strings, oversized tables).
- Refactors that tighten the HAL / SDK boundary.
- Pluggable per-device SDK layers (display, input, storage, battery) and per-device build configuration without forking the reader core.
- Documentation for adding a new ESP32 e-reader target.
- E-ink driver refinement (ghosting, partial update behavior).
Closed during this phase: new themes built into firmware, new external network connectors (sync engines, cloud storage, remote file access).
Goal: With the codebase smaller and portable, make reading great in every language: multi-language support, better font support with custom fonts, UI translations, and themes loaded from the SD card instead of consuming flash.
Focus areas:
- Multi-language reading support (underserved languages, complex script support where realistic on ESP32 hardware).
- Better font support and custom fonts.
- UI languages and localization.
- Moving themes off-firmware to SD-loaded assets (see SCOPE.md Section 6).
- Moving hyphenation files off-firmware. Hyphenation rules vary per language and the files are large (German alone is ~200KB). Today these eat flash budget that should be available for the reader core. The plan is to build a downloader analogous to the existing font downloader and store the dictionaries on SD / SPIFFS, loading on demand. This unlocks better hyphenation for long-word languages (German, Finnish, Norwegian, etc.) without paying the flash cost up front.
This phase depends on Phase 1 cleanup landing first; otherwise we generalize a moving target.
The following are explicitly not on the roadmap. They may live in other CrossPoint forks; they will not be picked up here:
- Interactive apps (games, calculators, notepads).
- Writing / authoring tools.
- Active connectivity features (RSS, news, browsers).
- PDF rendering as a first-class format.
See SCOPE.md for the full rationale.
- Phase boundaries are decided by maintainers, not by individual PRs.
- If a phase needs to be extended or an item carried over, that is documented here with a short note.
- Proposals for new phases or reordering should go through a Discussion first.