150 lines
11 KiB
Markdown
150 lines
11 KiB
Markdown
# VectorCraft Roadmap
|
||
|
||
**Stage: alpha** · next: beta, ~13 points (ready for real work ~62% → ~75%, and reliable `.ai` exchange) and ~170–270 h away
|
||
|
||
> **Last reviewed:** 2026-10-10 · **Last updated:** 2026-10-11 · **Change:** minor (Effect › Brush Strokes landed: Photoshop-style filters 28 of 57, raster effects 40% → 48%) · **Target:** Adobe Illustrator 2026 (30.x)
|
||
|
||
VectorCraft is a clean-room, open-source, pure-Rust reimplementation of the Adobe Illustrator workflow. It runs
|
||
on macOS, Windows, Linux, FreeBSD and the web (WASM), and agents can drive all of it over MCP, the CLI and a JSON
|
||
control channel. This page is the summary; the evidence is in [`docs/target-app-parity.md`](docs/target-app-parity.md)
|
||
and the work list in [`docs/gaps.md`](docs/gaps.md).
|
||
|
||
**Why alpha:** the core workflows exist end to end (drawing and path editing, Pathfinder, paint and appearance,
|
||
type, SVG/PDF/EPS, print, export) and feature depth is ~75%, but a professional can't yet switch: Illustrator's
|
||
own `.ai` files open with a wrong layer structure in real cases and can't be saved back with Illustrator's
|
||
editing data, the interaction details have never been checked side by side, and some machines fail to start the
|
||
app. Beta needs those closed (the beta list in [gaps.md](docs/gaps.md)). It passes the core-workflow
|
||
gate: all six core workflows (path drawing and editing, paint and style, type, layers/artboards/export, print
|
||
production, file exchange) work end to end on macOS with save and reopen, and only the last is partial, without
|
||
blocking it ([Alpha gate](docs/roadmap.md#alpha-gate)).
|
||
|
||
## Headline numbers
|
||
|
||
| | Value | Kind |
|
||
|---|---|---|
|
||
| **Feature breadth** (Illustrator's menu items, tools, effects, formats exist) | **~88%** | measured in part: menu items 348/375 wired (92.8%), tools 89/92, Illustrator effects 44/50, Photoshop effects 28/57, 3D 1/4 (initial Revolve) |
|
||
| **Feature depth** (weighted by use, scored by behaviour) | **~75%** | estimated |
|
||
| **To beta** | **~170–270 h** one agent · ~50–80 h with 4–6 agents | estimated |
|
||
| **To full parity** | **~360–590 h** one agent · ~95–165 h with 4–6 agents | estimated |
|
||
|
||
**Readiness by audience** (all estimated; method in [target-app-parity.md](docs/target-app-parity.md#ready-for-real-work)):
|
||
|
||
| Audience | Ready % | Opus 5.5 agent hours to ~95% (one agent) | With 4–6 agents | Work that dominates |
|
||
|---|---:|---:|---:|---|
|
||
| **Full Illustrator** (ready for real work; decides the stage) | **~62%** (57–67%) | **~340–560 h** | ~95–160 h | the mainstream work below, plus 3D (50–85 h), raster effects (24–40 h), advanced type (20–28 h), generative AI (30–60 h), localization (70–110 h) and ecosystem |
|
||
| **Mainstream illustrator** | **~65%** (60–70%) | **~230–380 h** | ~65–110 h | the interaction-fidelity pass (60–90 h), hardening and QA (30–50 h), stability (20–35 h), performance budgets (15–25 h), `.ai` exchange (10–20 h), the remaining depth of the everyday areas |
|
||
| **Essentials user** | **~75%** (70–80%) | **~60–105 h** | ~25–45 h | launch and stability on Windows (20–35 h), in-app help and onboarding (8–15 h), basic QA (10–15 h), Shape Builder regions (4–8 h), UI polish |
|
||
|
||
Hours are Opus 5.5 agent wall-clock hours, calibrated from this repo's merged-PR timestamps (e.g. 0.7–1.2 h per
|
||
Photoshop-style filter, ~0.8 h per medium feature); see [calibration](docs/target-app-parity.md#calibration).
|
||
The installed Illustrator was **not** inspected: the clean-room rule in [`AGENTS.md`](AGENTS.md) forbids it, so
|
||
everything is measured from our code against Illustrator's public documentation.
|
||
|
||
## By dimension
|
||
|
||
| Dimension | % | One agent (h) | Doc |
|
||
|---|---:|---:|---|
|
||
| Features (depth) | 75% | 220–360 | [target-app-parity.md](docs/target-app-parity.md) |
|
||
| UI/UX fidelity | 55% | 70–110 | [ui-parity.md](docs/ui-parity.md) |
|
||
| File formats | 75% | 15–30 | [file-format-parity.md](docs/file-format-parity.md) |
|
||
| Hardware (GPU, pen, touch, displays) | 35% | 25–45 | [hardware-parity.md](docs/hardware-parity.md) |
|
||
| Localization (12 key languages) | 54% (measured) | 70–110 | [localization-parity.md](docs/localization-parity.md) |
|
||
| Performance | ~60% | 15–25 | [gaps.md › G15](docs/gaps.md#g15-performance-budgets-and-a-gpu-renderer) |
|
||
| Stability | ~60% | 30–50 | [gaps.md › G10](docs/gaps.md#g10-launch-and-stability-on-real-machines) |
|
||
| Platforms | 80% | 15–25 | [gaps.md › G8](docs/gaps.md#g8-hardening-at-scale) |
|
||
| Ecosystem and plug-ins | 15% | 30–60 | [gaps.md › G17](docs/gaps.md#g17-ecosystem-third-party-plug-ins) |
|
||
| AI features (generative) | 0% | 30–60 + owner decision | [gaps.md › G18](docs/gaps.md#g18-generative-ai-features) |
|
||
| Agent automation | beyond Illustrator | — | [mcp.md](docs/mcp.md) |
|
||
|
||
The dimension hours overlap (feature rows hold some UI, format and AI work), so they don't add up to the total.
|
||
|
||
## Features
|
||
|
||
| Area | % | One agent (h) |
|
||
|---|---:|---:|
|
||
| Selection, transform & align | 85% | 4–6 |
|
||
| Drawing tools | 85% | 8–13 |
|
||
| Path operations, Pathfinder, Shape Builder, Live Paint | 80% | 4–8 |
|
||
| Colour, swatches, gradients, patterns, mesh, recolor | 92% | 4–7 |
|
||
| Strokes, brushes, width profiles | 88% | 8–15 |
|
||
| Appearance, transparency, graphic styles, masks | 97% | 1–2 |
|
||
| Live vector effects | 85% | 6–10 |
|
||
| Raster effects (Effect Gallery) — [effects-parity.md](docs/effects-parity.md) | 48% | 24–40 |
|
||
| 3D and Materials | ~8% | 50–85 |
|
||
| Type core — [type-parity.md](docs/type-parity.md) | 80% | 14–19 |
|
||
| Type advanced (CJK, spelling, Touch Type) | 50% | 20–28 |
|
||
| Symbols, blends, envelopes, Repeat, perspective | 85% | 6–10 |
|
||
| Image Trace, graphs, image tools | 75% | 5–8 |
|
||
| Layers, artboards, document setup | 78% | 5–8 |
|
||
| View & navigation | 70% | 9–14 |
|
||
| Guides, grids, smart guides, snapping | 80% | 4–7 |
|
||
| File formats — [file-format-parity.md](docs/file-format-parity.md) | 75% | 10–20 |
|
||
| Export for Screens, Asset Export, slices | 85% | 2–4 |
|
||
| Print, colour management, separations | 80% | 2–4 |
|
||
| Automation (Actions, Variables, scripting) | 60% | 8–14 |
|
||
| UI chrome (panels, Properties, preferences) | 80% | 13–20 |
|
||
| Libraries, Links, Package | 82% | 2–3 |
|
||
| Generative AI | 0% | 30–60 |
|
||
|
||
Weights, evidence and the missing items per area: [target-app-parity.md](docs/target-app-parity.md#feature-areas).
|
||
|
||
## Languages
|
||
|
||
| Language | Code | Status | Translated |
|
||
|---|---|---|---:|
|
||
| English | `en` | full | 100% |
|
||
| Simplified Chinese | `zh-hans` | partial (messages in English) | 85% |
|
||
| Spanish | `es` | full | 100% |
|
||
| Hindi | `hi` | none | 0% |
|
||
| Arabic | `ar` | none (no right-to-left interface) | 0% |
|
||
| French | `fr` | full | 100% |
|
||
| Portuguese (Brazil) | `pt-br` | partial | 68% |
|
||
| Indonesian | `id` | none | 0% |
|
||
| Japanese | `ja` | full | 100% |
|
||
| German | `de` | full | 98% |
|
||
| Korean | `ko` | none | 0% |
|
||
| Vietnamese | `vi` | none | 0% |
|
||
|
||
Also shipped: Traditional Chinese (partial, 80%), Italian, Russian and Ukrainian (full), Czech (menus only).
|
||
No catalog has had a native speaker's review. Details: [localization-parity.md](docs/localization-parity.md).
|
||
|
||
## Upcoming
|
||
|
||
Ranked; detail in [docs/roadmap.md](docs/roadmap.md) and [docs/gaps.md](docs/gaps.md).
|
||
|
||
1. **Interaction-fidelity pass** (G1), tool by tool from [ui-parity.md](docs/ui-parity.md): 60–90 h.
|
||
2. **`.ai` files from real users** (G9): layer structure, guides, speed: 10–20 h.
|
||
3. **Launch and stability on real machines** (G10): 20–35 h.
|
||
4. **Core-tool correctness** (G11: Shape Builder regions, Asset Export borders): 8–15 h.
|
||
5. **Raster effects** (G2) alongside: 29 filters and the Effect Gallery, 24–40 h.
|
||
6. **Then:** type core and advanced type (G4), the Illustrator 2026 additions (G13), 3D (G3).
|
||
|
||
## Progress log
|
||
|
||
| Date | What landed |
|
||
|---|---|
|
||
| 2026-10-11 | Effect › Brush Strokes: Accented Edges, Angled Strokes, Crosshatch, Dark Strokes, Ink Outlines, Spatter, Sprayed Strokes, Sumi-e (G2). Command parameters from agents and files checked before anything changes (paint, place, cuts, dashes, crop, asset ids, anchors, paths, distort; #984–#999); command dialogs keep whole numbers whole, so Create Object Mosaic takes the tiles typed (#1000); stepper arrows take Shift (×10) and Ctrl/Cmd (÷10) like the arrow keys (#991). Exports and Rasterize supersample Art Optimized edges, so shapes meeting edge to edge leave no seam (#983). Character and paragraph styles take leading, kerning, mojikumi and the other attributes left out at their defaults (#1001). |
|
||
| 2026-10-10 | Initial live Revolve (Effect › 3D and Materials, #846, #605), in its own `three-d` crate. Progress docs re-measured and restructured to the craftrules standard (this page, `docs/target-app-parity.md`, `gaps.md`, `ui-parity.md`, the format, hardware, localization, effects and type checklists). Same day: v0.8.0; Effect › Distort, Pixelate and Texture filters (13); German interface; Variables (data merge); Graph Type value axes and tick marks; restart when the first frame never reaches the screen (#964) |
|
||
| 2026-10-09 | Radial Blur, Smart Blur, Unsharp Mask (gap 2 begins); 340 commits, the busiest day |
|
||
| 2026-10-08 | v0.5.0–v0.7.0; community contributions: Pen and shape-tool modifiers, Layers, Artboards, Japanese composition, Hebrew/Arabic type, interface languages, saved selections |
|
||
| 2026-10-07 | v0.4.0; issue fixes; Layers panel rework |
|
||
| 2026-10-06 | M8 transform and distort fidelity pass (M8.1–M8.20) |
|
||
| 2026-10-05 | M4 files done (M4.14–M4.98); M3 paint and appearance done (92 tasks over two days) |
|
||
| 2026-10-02 | First weighted parity estimate (22 areas); look and feel measured against public screenshots |
|
||
| 2026-10-01 | Renamed from DrawCraft to VectorCraft |
|
||
| 2026-09-30 | Repository history begins |
|
||
|
||
The full pre-standard record of what landed is kept in [docs/roadmap.md](docs/roadmap.md#progress-log-before-the-progress-docs-standard).
|
||
|
||
## Revision history
|
||
|
||
| Date | Change | Summary |
|
||
|---|---|---|
|
||
| 2026-10-10 | minor | Method aligned with the standard, no new evidence: full ready for real work is the additive weighted sum, ~62%; hours per audience; beta ~13 points away |
|
||
| 2026-10-10 | minor | Added the essentials-user readiness score (~75%) |
|
||
| 2026-10-10 | minor | Ready for real work re-examined against the issue tracker and user feedback: 50% → 55% (full), 65% mainstream added; see target-app-parity.md |
|
||
| 2026-10-10 | minor | Applied the core-workflow gate (docs/roadmap.md › Alpha gate): passes, stage stays alpha |
|
||
| 2026-10-10 | major | Re-measured against Illustrator 2026 (30.x) from code counts, 71 open issues and public docs; stage set to alpha; moved the parity estimate, honest assessment and "Shipped so far" to `docs/target-app-parity.md`, the gap list to `docs/gaps.md`, milestones to `docs/roadmap.md`, the `.ai` scope to `docs/file-format-parity.md` |
|
||
| 2026-10-05 | major | Honest assessment by dimension and the prioritized gap list |
|
||
| 2026-10-02 | major | First 22-area weighted parity estimate |
|