Phase Plan (P0-P5)

Snapshot: 2026-05-18 Canonical schedule source: 47-gtm-plan — submission target NLT 2026-07-09. Live status sources: 11-weekly-updates for sprint closeouts · 32-tb11-bug-tracker for bug inventory. Scope of this document: execution phase structure, gate definitions, workstreams, and decision gates. Schedule and live-status claims have been removed from this revision — refer to the canonical sources above.

Rebaseline (2026-05-18 snapshot): ADN target pulled to NLT 2026-07-09. See 47-gtm-plan for current schedule. Hard gate rules below stand regardless of date drift.

Execution Authority

This document is the canonical execution and gate plan. Strategic authority is 00-vision-and-success-criteria.md; canonical schedule is 47-gtm-plan.md. If this document conflicts with 00 or 47, update this document to match.

Program execution truth

  1. Primary release path remains Autodesk App Store .bundle.

  2. WiX/MSI is contingency-only and not the active critical path.

  3. Active critical path dependency is progcont routing reconstruction, comprehensive parity validation (45-tb11-parity-test-plan), and validation completion.

  4. 37-cv-web-vs-pascal-editor-comparison-strategy.md defines recommended web/editor architecture direction and non-goals for 2026.

Gate alignment

  1. P0 remains complete historically, but active validation closure work continues under P1 execution control.

  2. P1 is considered active until required validation and release-gate artifacts are complete.

  3. P2+ phase start dates are dependent on P1 gate closure, not fixed calendar-only assumptions.

Milestone arc

Current milestone windows: see 47-gtm-plan. Hard gate rule (below) stands regardless of date drift.

Hard gate: M4 security testing cannot start until all validation steps are complete in active configs, including Step 0-12 closure, regression reruns, and DFMEA-linked high severity findings.

Cross-document dependencies

  1. 04-2026-timeline-weekly.md covers historical week cadence; current sprint sequencing is in doc 47.

  2. 10-milestones-dashboard.md is historical-arc; current milestone status is in doc 47.

  3. 07-release-and-distribution.md owns release-content structure; current sprint sequence is in doc 47.

  4. 05-risk-register.md owns the full risk register.

Overview

Your weekly status definitions are the backbone. This plan translates them into:

  • explicit deliverables

  • decision gates

  • staffing assumptions

  • workstream structure


P0 — VM “Run as-is” (DONE)

Status: ✅ Completed 2026-01-13, Upgraded 2026-01-27 Definition: Runs on VM on new machine; dry run successful in AutoCAD 2000. Constraints: Not distributable without selling laptop or custom install; version may be behind newest build.

Deliverables

  • Captured VM image + backup

  • “Golden run” demo recording and checklist

  • Initial inventory notes (what features matter most)

P0 Infrastructure Upgrade (Jan 27, 2026)

ConstructiVision Tower → Proxmox VM Migration:

  • ✅ Tower computer converted to Proxmox VM (only running legacy copy)

  • ✅ Properly backed up with snapshots (can clone/revert safely)

  • ✅ Tailscale VPN configured for secure remote access

  • ✅ LegacyUpdate.net integrated (keeps XP/Vista patched)

  • ✅ Greater Seattle Concrete remote access in progress

Benefits over VirtualBox on Dev Laptop:

Aspect

Before (VirtualBox)

After (Proxmox)

Location

Dev laptop (local)

Dedicated server

Backup

Manual, infrequent

Automated snapshots

Resources

Shared with dev work

Dedicated

Remote Access

None

Tailscale VPN

Customer Testing

Not possible

GSC can access directly


P1 — Distribution Modernization (App Store First)

Status: InstallShield and WiX tracks aborted for primary distribution. Key Finding: Existing VLX/runtime path can be packaged for .bundle delivery with lower complexity than MSI. Decision: Primary path is Autodesk App Store .bundle; MSI is contingency-only for enterprise exceptions.

P1 Pivot Rationale

  • Low ROI: .rul recompilation had issues, wasn’t going smoothly

  • Better Path Found: Total Uninstall backup/restore validated on Win10x32

  • Auto-Config Script: scripts/Configure-ConstructiVision.ps1 handles AutoCAD integration

  • Result: Distribution can proceed through App Store .bundle packaging without resurrecting legacy installer toolchains.

Current Approach: Total Uninstall + .bundle

  1. File Payload: Total Uninstall inventory (152 files, 5.5 MB)

  2. Registry Config: AutoCAD profile settings via PowerShell script

  3. Submission assets: listing metadata, legal/policy artifacts, support links, and release evidence

Platform Compatibility (Validated Feb 13)

Platform

AutoCAD 2000

ConstructiVision

Status

Windows XP

✅ Works

✅ Works

Fully Supported

Windows Vista

✅ Works

✅ Works

Fully Supported

Windows 7

✅ Works

⚠️ BHF bug

Limited

Win10 32-bit

✅ Works

✅ Works

Fully Supported

Win10 64-bit

✅ Runs

✅ Works (registry fix)

Supported — requires Wow6432Node COM fix

Phase 1A: Analysis & Setup (✅ COMPLETE)

Completed: January 21 – February 24, 2026 Focus: Environment baseline, legacy analysis, architecture decisions

Key Discoveries:

  1. isDcc decompilation success (Completed 2026-01-20)

    • Tool: isDcc v1.22 (Luca Borsato)

    • Input: setup.ins (compiled InstallScript)

    • Output: setup.rul (4,163 lines of readable InstallScript)

    • Now have full source for installer logic

  2. ⚠️ VLX/FAS cannot be decompiled

    • Compiled Visual LISP format has no reliable decompiler

    • Solution: Recompile from LSP source files using VLIDE

    • All .lsp source files present in v7.0(patch)/ directory

Analysis Workstreams (historical):

  • Week 1 (Jan 21-27) — Legacy Environment Baseline: InstallShield 5.x in XP VM, setup.rul import, AutoCAD detection bug documented, baseline build environment established. PE Header Analysis confirmed CV SETUP.EXE is 32-bit but AC2000 SETUP.EXE is 16-bit NE (blocks Win11 fresh installs).

  • Week 2 (Jan 27 – Feb 3) — Gap Analysis: v7.0(patch) inventory, v3.60 vs v7.0 differences, unified source inventory, missing components.

  • Week 3 (Feb 3-10) — Nero Analysis: isDcc bytecode output (10,031 lines) won’t compile; wizard approach generates clean 606-line .rul. Decompiler useful for understanding, not rebuilding.

  • Week 4 (Feb 10-16) — Multi-OS Testing: AC2000 + CV3.60 working on XP, Vista, Win7. 8 bugs documented with root cause analysis.

  • Week 5 (Feb 10-17) — Win10 Testing: Win10 32-bit + 64-bit working. Auto-configuration script created. Total Uninstall restore validated.

  • Week 5b (Feb 16) — v3.60 Source Recovery: 12 modules recovered → merged into TB11-01x32, 3 crash paths fixed, 9 silent startup load failures resolved. Nightly deployment pipeline (SSH deploy keys + sparse checkout) on VMs 108, 109, 201, 202.

Feb 13 Decision: P1 InstallShield Sprint ABORTED

  • Low ROI on .rul recompilation

  • Total Uninstall payload + auto-config script = complete installer components

  • New Path: Proceed directly to App Store .bundle packaging using validated components

Subsequent execution roadmap: historical Jan–Mar 2026 week-cadence detail is preserved in 04-2026-timeline-weekly.md; current sprint sequencing lives in 47-gtm-plan.

Completed Prerequisites:

  1. Build environment reproduction (Completed 2026-01-16) — v3.60 with v11 hotfix captured; deploy-ready packages archived in Git LFS.

  2. Legacy installer analysis (Completed 2026-01-20) — InstallShield 5.0 installer decompiled via isDcc v1.22; AutoCAD detection bug, menu registration bug, and Vista+ incompatibility documented. setup.rul analysis captured for archive; active distribution path is .bundle packaging.

  3. VM-based interim deployment (Upgraded 2026-01-27) — Proxmox VM server replaces dev-laptop VirtualBox; Tailscale VPN + LegacyUpdate.net; GSC remote access enabled; first-time proper snapshot backup.

Phase 1B: Distribution Development (🚧 .bundle pivot)

Timeline: see 47-gtm-plan §12-week schedule. Framework: WiX Toolset 4.xVLX .bundle format for Autodesk App Store Prerequisites: ✅ Total Uninstall payload validated, ✅ Auto-config script tested Status: WiX ABORTED (Feb 28). Pivoted to .bundle packaging.

Why .bundle? (Feb 28 Decision)

  • WiX development started but distribution strategy was re-evaluated

  • .bundle format is natively supported by Autodesk App Store

  • Provides simpler deployment model than MSI for AutoCAD plugins

  • Aligns with Autodesk’s preferred plugin distribution model

  • Reduces installer complexity — no need for MSI, registry manipulation handled by .bundle loader

P0 Automated Validation Sprint (Mar 1-4, COMPLETED):

Before building .bundle packaging, systematic validation of TB11 source-mode correctness was executed:

  • AutoIT Automation: cv-p0-step0.au3 (launch), cv-p0-step1.au3 (layout), cv-p0-step2.au3 (CSV init — 11 tests)

  • OCR Pipeline: scripts/ocr-screenshots.py for automated screenshot comparison against golden baseline

  • 6-Config Test Matrix: Configs A-F across VMs 102, 103, 104, 108, 109

  • P0 Steps 0-2 PASS on Configs A-E (5/6)

  • Config F (VM 109 x64) FAIL — Bug 38 (ACAD path → stale acad2000 subdir). Fixed cecc9ef.

  • 18 bugs found and 17 fixed during validation sprint (Bugs 22-38)

  • Win10 PB11 ABANDONED (Bug 37) — WinHelp() in VLX incompatible with Win10. TB11 source-mode only for Win10.

  • 🚧 VLX/source progcont mismatch discovered — CSV.VLX was compiled from source not in repo

Components Ready:

  1. File Payload — Total Uninstall inventory validated

  2. Auto-Config Scriptscripts/Configure-ConstructiVision.ps1 tested (incl. x64 auto-detect + acad2000 cleanup)

  3. TB11 Source-Mode — P0 Steps 0-2 validated on 5/6 VM configs

  4. Serialization — License key validation (TBD)

  5. .bundle Packaging — sprint scheduling in doc 47

Remaining Workstreams (sprint sequencing in doc 47):

  1. P0 Validation Completion — Retest VM 109 (Config F) after Bug 38 fix, P0 Step 3 (open CSB001.dwg), P0 Step 4+ menu command testing, progcont routing reconstruction.

  2. .bundle Packaging — structure prototype, PackageContents.xml manifest, Autoloader integration, Win10 x32/x64 testing.

  3. Alpha Testing & Release Gate Closure — internal test VM deployment, deployment guide, release candidate build.

Phase 1E: Comprehensive Parity Validation (Doc 45)

Timeline: overlaps with Phase 1B/.bundle packaging — see 47-gtm-plan §blockers. Canonical plan: 45-tb11-parity-test-plan.md Goal: Verify TB11 source-mode parity with PB11 VLX-mode across all user-facing functionality

Workstreams (8 phases, 92 tests):

  1. Phase 1 — Menu Routing (G1): 17 progcont paths reach correct dialogs

  2. Phase 2 — Dialog Appearance (G2): Visual parity with PB11 golden screenshots

  3. Phase 3 — Field Functionality (G3): Input, toggle, dropdown, calculation validation

  4. Phase 4 — Data Persistence (G4): NOD XRecord read/write round-trip

  5. Phase 5 — Drawing Operations (G5): Panel creation, site drawing, batch operations

  6. Phase 6 — Print & Output (G6): Print preview, reports, CSV/DXF export

  7. Phase 7 — Infrastructure (G7): Startup, SFSP, profile, deployment

  8. Phase 8 — Regression: Full sweep after all fixes applied

Exit criteria: ≥95% of 76 testable items PASS/PARTIAL. G1 and G4 are zero-FAIL gates. 12 items are BLOCKED (pending reconstruction backlog), 4 DEFERRED. Master scorecard in doc 45 Section 11. DFMEA-036 through DFMEA-042 define the new failure modes.

Gate dependency: M2 cannot close until Phase 1E exit criteria are satisfied.

Phase 1D: Integration & Bug Fixing (⏳ NEW PHASE)

Timeline: see 47-gtm-plan §12-week schedule. Goal: Validate old ConstructiVision works with modern AutoCAD; fix critical bugs; establish stable baseline Focus: Functional modernization only (no UI changes)

Workstreams (sprint sequencing in doc 47):

  • AutoCAD 2020+ compatibility testing — test old ConstructiVision VLX with AutoCAD 2020/2022/2024, document compatibility issues and breaking changes, identify required code changes for modern AutoCAD API.

  • Functional modernization (no UI changes) — update deprecated AutoCAD API calls, fix compatibility issues, maintain existing UI/UX (defer visual changes to P2), LSP code updates for modern AutoCAD.

  • Bug fixing sprint — address critical bugs found during multi-OS testing, fix hard errors from menu validation testing, regression testing on XP/Vista/Win7/Win10, AutoCAD version-specific fixes.

  • Stabilization & testing — integration testing, performance testing and optimization, documentation updates, final regression testing across all supported platforms. Decision point: stable baseline → proceed to P2 Design.

Phase 1C: OS/AutoCAD Compatibility Validation (✅ COMPLETE Feb 13)

Status: COMPLETE — Full compatibility matrix validated

OS Testing Matrix:

OS

Priority

VM

Status

Results

Windows XP SP3

P1

VM 104

PASSED

AC2000 + CV3.60 working

Windows Vista SP2

P2

VM 106

PASSED

No UAC issues as Administrator

Windows 7 SP1

P2

VM 107

⚠️ LIMITED

pcAnywhere BHF path issue (not a CV bug — see doc 22)

Windows 10 32-bit

P1

VM 108

PASSED

Full stack working with auto-config script

Windows 10 64-bit

P1

VM 109

PASSED

Wow6432Node COM fix applied — fully working

Windows 11

P3

N/A

🚫 Deferred

Need 64-bit AutoCAD first

Win10 64-bit Resolution (Feb 17):

  • Total Uninstall restore bypassed 16-bit installer; AutoCAD 2000 ran on 64-bit Windows; CV files loaded successfully.

  • Resolved: CSV command worked after Wow6432Node COM registry fix.

  • Root Cause: Total Uninstall placed COM registrations in native 64-bit hive instead of Wow6432Node.

  • Fix: Copy 103 CLSIDs, 104 IAcad Interfaces, and 6 ProgIDs to Wow6432Node (see 27-vm109-autocad-x64-registry-fix.md (sensitive)).

  • Additional: Project Files search path must be manually set in AutoCAD Options.

AutoCAD Testing Matrix:

AutoCAD Group

Versions

Architecture

VLX Support

Legacy 32-bit

R14, 2000-2006

32-bit only

Original VLX

Transition

2007-2009

32-bit

VLX compatible

Modern 32/64

2010-2013

Both

Test both

64-bit Only

2014-2026

64-bit only

May need rebuild

Exit criteria for P1

  • Analysis Exit: Legacy environment fully baselined; VLX files recompiled from LSP source; tested on XP, Win 7, Win 10, Win 11 path documented; compatibility matrix complete; requirements locked for Installer phase.

  • Installer Exit: WiX installer GA release → superseded by App Store .bundle path.

  • Documentation Exit: Remote access working for beta customers (VM or direct install); screenshot tree complete for UI modernization planning; failure codes and limitations documented; IT deployment guide complete.

Decision gate into P2 (Design Improvement)

  • Is legacy environment fully baselined?

  • Which installer approach should be primary (App Store .bundle)?

  • Are beta customers satisfied with installation experience?

  • Is VLX recompilation sufficient or do we need source modifications?


P2 — 64-bit AutoCAD + Alpha/Beta Testing

Status: P2 active. Goal: Internal testing, bug fixing, and output quality validation before customer release.

P2-A: 64-bit AutoCAD Compatibility

Status: AC2000 x64 working via Wow6432Node registry fix.

  • ✅ AC2000 working on x64; VLX loads and runs on x64; no recompilation needed.

  • Existing VLX works with AutoCAD 2000 on 64-bit Windows.

  • COM automation works on 64-bit Windows after Wow6432Node registry fix.

  • VLX recompilation is not needed for x64 — identical binary works on both architectures.

  • Modern AutoCAD (2010+) pursued in P3 for future-proofing.

P2-B: Alpha/Beta Testing

Alpha testing with GSCI engineers began Feb 17, 2026 on isolated Win10 32-bit VMs (200, 201) cloned from VM 109. Tailscale PVE authentication + Remote Desktop. Current alpha status: see 11-weekly-updates.

See: alpha-testing-gsci-plan.md (sensitive) for complete details · alpha-testing-gsci-checklist.md (sensitive) for the checklist.

P2-C: Customer Validation

  • Deploy working installer to Greater Seattle Concrete

  • Gather real-world feedback

  • Iterate based on findings

P2-D: Blueprint Extraction R&D (Track D) — pulled into v11.1 ship gate

Status (2026-05-18): Track D was previously framed as post-submission R&D. It has been pulled into the v11.1 ship gate so the cv-cad App Store submission ships with both a working extraction prototype and patent-evidence figures. Stories CV-326 through CV-336 land in Sprints 21–25 (P2 window); CV-337 + CV-338 land in Sprints 26–27 (P3 window).

  • Tier 1 (already wired, Sprint 19): pdf-mcp + ocr-mcp MCP servers.

  • Tier 2 (Sprints 21–24): Local VLM baseline — Qwen2.5-VL / Qwen3-VL 8B zero-shot on fixture PDFs; bbox quality, symbol P/R, end-to-end pipeline test, Tier 2→Tier 3 decision.

  • Tier 3 (Sprints 21–27): cv-blueprint-mcp custom MCP — .cvextract.json schema v0.1 ships Sprint 21 (CV-336), scaffold + 3 tools Sprint 22 (CV-332), vector path Sprint 23, raster + scale calibration Sprint 25.

See 47-gtm-plan §Track D for the full sprint-level schedule and 50-blueprint-extraction-architecture for the architecture spec.


P3 — Security, Testing, and Pre-Submission Prep

Status: Sequenced against the NLT 2026-07-09 submission target — see 47-gtm-plan §12-week schedule. Goal: Security hardening, multi-version compatibility testing, ADN/App Store prep. Dependency: AutoCAD 2026 access for validation and packaging.

P3-A: UI/UX modernization (deferred to v1.1 per doc 47 §scope compression)

  • Review all DCL dialogs for usability improvements

  • Modernize outdated interface patterns

  • Improve error messages and user feedback

  • Update visual styling where possible within DCL constraints

UI/UX modernization scope has been deferred out of the v1.0 submission window. See doc 47 for the v1.0/v1.1 scope split.

P3-B: Code Refactoring

  • Clean up legacy LSP code structure

  • Improve maintainability and documentation

  • Remove dead code and redundant logic

  • Standardize naming conventions and patterns

P3-C: Documentation

  • Complete user guide with screenshots

  • Installation and configuration guide

  • IT deployment guide for enterprise

  • Developer documentation for future maintenance

P3-D: Security & Testing Workstreams

  1. Multi-Version Compatibility Testing

    • Test on AutoCAD 2024, 2025, 2026

    • AutoCAD API compatibility fixes (LISP/VLISP behavior differences)

    • Performance pass + startup reliability

    • Build comprehensive test matrix

  2. Security Audit & Hardening

    • Code security review (input validation, file handling)

    • Vulnerability scanning (static analysis)

    • Dependency audit (third-party components)

    • Windows Defender / SmartScreen compatibility

  3. Autodesk Developer Network (ADN) Preparation

    • Apply for ADN Developer membership (~$500/year)

    • Request early access to AutoCAD 2027 SDK

    • Review App Store submission guidelines

    • Prepare compliance checklist (security, EULA, privacy policy)

  4. Modern Plugin Format (.bundle Migration)

    • Convert from acad.rx loading to .bundle structure

    • Enables Autodesk App Store compatibility

    • Improves Windows 11 security compliance

P3-E: Blueprint Extraction Patent-Evidence Checkpoint (Track D, Sprints 26–27)

The Track D pipeline closes the v11.1 ship gate in P3 with two patent-evidence deliverables:

  • CV-337 (Sprint 26) — cv-cad round-trip. First end-to-end scanned plan .cvextract.json cv-cad TB11 input import demo on a reference project. This is the patent-evidence checkpoint — the figure the attorney cites for the “extracting structured measurements from construction drawings” claim. Lands one week before the NLT 2026-07-09 ADN submission deadline.

  • CV-338 (Sprint 27) — competitive baseline. Benchmark cv-blueprint-mcp vs LandingAI ADE on 10 representative drawings. Report attached to the ADN submission package as differentiation evidence.

See 47-gtm-plan §Track D and 50-blueprint-extraction-architecture. Ship-gate risks R26/R27/R28 (00-vision-and-success-criteria §Top Risks).

Exit criteria

  • Clean load and run on AutoCAD 2024, 2025, 2026

  • Security audit passed with no critical findings

  • Regression suite passes on all target environments

  • .bundle format validated and working

  • ADN Developer membership active

  • Ready for code signing and final packaging

Decision gate into P4

  • Is multi-version compatibility solid enough for production?

  • Are all security findings addressed?

  • Do we need ADN Partner tier ($2K-10K) or is Developer tier sufficient?


P4 — Final Release & Demo

Status: Sequenced against the NLT 2026-07-09 submission target — see 47-gtm-plan §12-week schedule. Goal: Code signing, App Store submission, owner demo, and production release. New Objectives: Dual distribution (direct download + Autodesk App Store)

Workstreams

  1. Code Signing & Packaging

    • Obtain EV (or OV) code signing certificate

    • Sign .bundle package for SmartScreen trust

    • Create final distributable packages

    • Validate on clean Windows installations

  2. Autodesk App Store Submission

    • Submit to Autodesk App Store

    • Address security review findings (allow 2 cycles)

    • Publish listing (screenshots, description, pricing)

    • Monitor downloads and conversion rates

  3. Direct Distribution Channel

    • Download portal on website

    • Auto-update mechanism

    • IT deployment guide (Group Policy, SCCM)

    • Contingency: enterprise MSI path documented in 07-release-and-distribution.md §Contingency.

  4. Owner Demo & Handoff

    • Prepare demonstration environment

    • Create handoff presentation

    • Document operational procedures

    • Knowledge transfer session

  5. Web Developer Onboarding (post-revenue 2027)

    • Hire web developer

    • Begin AutoLISP → .js/.NET conversion planning

    • Set up development environment

    • Initial prototype work

Distribution Strategy:

Channel

Pros

Cons

Target Audience

Direct Download

Full control, no revenue share

Must market ourselves

Existing customers, referrals

Autodesk App Store

Trusted marketplace, wider reach

30% revenue share to Autodesk

New customers, enterprise IT

Exit criteria

  • Code signing certificate obtained and applied

  • .bundle package signed and validated

  • Autodesk App Store listing approved and live (or in review)

  • Direct download portal live and tested

  • Owner demo completed successfully

  • 🎉 ConstructiVision v26 Released

Decision gate for future (P5)

  • Which distribution channel performs better (direct vs App Store)?

  • Should we invest in ADN Partner tier for co-marketing?

  • Is web conversion viable as primary product direction?


P5 – Web Deployment & AI Enhancements (2027+)

Status: TBD Goal: Web-based access + AI-assisted automation Start: Begins post-submission with web developer hire; full deployment 2027+

Web Deployment Strategy

Why Web Deployment?

  • Removes AutoCAD desktop dependency for some workflows

  • Enables broader market reach (non-AutoCAD users)

  • Supports mobile/tablet access

  • Enables SaaS subscription model

Technology Options:

Option

Pros

Cons

Effort

Best For

JavaScript (Browser)

Universal access, no install

No native AutoCAD integration

Medium

Standalone tools

APS Design Automation

Official Autodesk API, headless AutoCAD

Per-API-call costs, learning curve

High

Heavy CAD processing

.NET Web API

Strong typing, familiar stack

Requires hosting, no AutoCAD runtime

Medium

Backend services

Recommended Phased Approach:

  1. Phase 5A: JavaScript utilities (form fillers, calculators) - Low risk

  2. Phase 5B: .NET API backend (data services, report generation) - Medium risk

  3. Phase 5C: APS Design Automation (CAD automation as service) - High investment

APS (Autodesk Platform Services) Integration:

  • Design Automation API for headless AutoCAD execution

  • Data Management API for file storage/version control

  • Model Derivative API for viewing without AutoCAD

  • Note: Requires APS subscription (separate from ADN)

AI Enhancements

Deliverables (incremental)

  • AI-assisted form fill (structured extraction + review UI)

  • Panel book generator “EZ Button” (template + rules + traceability)

  • Integrated test output generation (evidence pack per project)

Exit criteria

  • Human-in-the-loop review workflow

  • Audit-friendly traceability (what data → what output)

  • Clear off-switch and fallbacks


Security & Compliance Requirements

Code Signing (Required for P1+):

Certificate Type

Cost/Year

SmartScreen

Purpose

Standard Code Signing

$200-400

Builds trust slowly

Development/testing

EV Code Signing

$400-700

Immediate trust

Production release

Recommended Vendors: DigiCert, Sectigo, Certum

Autodesk Developer Network (Required for P3+):

Tier

Cost/Year

Benefits

Developer

~$500

App Store access, basic support

Partner

$2K-10K

Early SDK access, co-marketing

App Store Compliance Checklist:

  • Security review (automated scan + manual review)

  • EULA approved by legal

  • Privacy policy (data collection disclosure)

  • Terms of service

  • .bundle format validated

  • Screenshots and marketing materials


Staffing notes (levers that move the schedule)

  • Core engineer (AutoCAD/LISP/VLISP): required for P1–P3

  • Release engineer / installer: accelerates P1–P3 and reduces support risk

  • Web developer: enables P4 parallelization and earlier P5 starts

  • QA / regression automation: reduces parity risk and accelerates confidence