nobody should be priced out of dignity
OpenSourceRail is an open urban-rail reference platform for countries that want to retain design authority, software, fabrication, integration, operations and maintenance capability in-country. It connects city planning, GIS, CAD/IFC, simulation, project controls, local manufacturing, ERP, equipment supervision and assurance around the same city, asset and engineering-revision identities.
Important
Repository outputs are planning and engineering-screening evidence—not bids, construction releases, safety certificates, approvals or endorsements. Open topography and water data do not replace survey, geotechnical or hydraulic evidence. Simulation and formal checks do not authorize live railway command.
Start here: two-page brochure · complete PDF book · one-page overview · architecture · current roadmap and open work
Native part generation: generate drilled steel cleats and RHS/I-section civil beams through FreeCAD → Gmsh → CalculiX, then use native feedback to choose subsequent designs. Deterministic bounded search retains failed cases, provenance, geometry and solver files. Best-design claims cover the declared, fully tested numerical domain; physical qualification remains open. Run the recursive workflow · Review twelve related design systems. OpenSourceRail connects corridor choices to engineering revisions, local production, delivery costs and maintenance evidence. Its planning portfolio covers 265 cities in 44 developing countries, plus one European comparison model. This table compares published functions, not performance, price or railway acceptance. Reviewed 12 October 2026: linked primary sources; capabilities depend on edition/configuration. “Integrated” means an existing OSR component; “workflow” means a tool used with OSR; “benchmark” implies no connector.
| Tool / suite | Availability / relationship | Main functions | OpenSourceRail's corresponding value and limits |
|---|---|---|---|
| OpenSourceRail | Open source; connected reference platform | City GIS, service/energy simulation, CAD/IFC, quantities, local manufacture, finite-resource CPM, procurement/cashflow, ERP, supervision and maintenance evidence | Common city/asset/revision identities connect planning through repair; reproducible demonstrations exist, while physical railway qualification and live control acceptance remain open. |
| Siemens TPS.plan, TPS.live, TPS.trackworks | Commercial; benchmark | Infrastructure-aware timetables, traffic forecasts/conflicts, dispatching and track-work windows | OSR schedules declared physical routes around possessions and resource conflicts, with fleet/energy and staff recovery; search is bounded and live dispatch acceptance remains open. |
| Alstom Urbalis Vision / ICONIS supervision | Commercial; benchmark | Integrated urban traffic supervision, regulation, SCADA, incident and depot management | OSR joins OCC, simulator and observation-only FUXA views to asset/work records; its management stack is not an accepted live ATS/CBTC product. |
| Alstom HealthHub / FlexCare digital services | Commercial; benchmark | Fleet, infrastructure and signalling condition monitoring, predictive maintenance and service support | OSR joins condition-to-repair and serial/parts history to exposure/fault trends and held-out degradation/RUL validation; measured calibration and predictive deployment remain open. |
| Hitachi Rail HMAX | Commercial; benchmark | Digital asset management, harmonised health data, condition/predictive maintenance and digital twins | OSR exposes engineering provenance and condition-to-repair evidence in an adaptable owner-held stack; deployed predictive depth and support equivalence are unproven. |
| Hitachi Energy EAM / FSM / APM | Commercial; benchmark | Rail asset registers, track defects, work orders, field service and asset performance | OSR links assets, work and independent handback to revisions, with offline typed/photo work packs, durable synchronisation and native-state revalidation; operator commissioning remains open. |
| Dassault DELMIA Quintiq | Commercial; benchmark | Coupled service, fleet and crew planning, constraint optimisation and disruption replanning | OSR proposes possession-aware service/fleet/energy recovery and bounded staff rosters, retaining deadheads, charging and fixed maintenance; unrestricted optimisation depth remains unproven. |
| Tracsis ATTUne / TRACS / TRACS Enterprise | Commercial; benchmark | Timetables, vehicle/crew diagrams, rostering, personnel and on-day operations | OSR proposes rosters across operating and maintenance roles, checking competence, travel, rest, breaks and workload; native assignment and deployed rostering acceptance remain separate. |
| Wabtec RailConnect | Commercial; freight benchmark | Railcar movement, switching, mobile work, freight revenue/billing and industry exchanges | OSR's current reference is urban passenger rail with engineering-to-maintenance continuity; freight settlement and car-hire workflows would require a separate profile. |
| Petrosoft RAILSoft / mRails | Commercial; benchmark | Fleet planning, maintenance intervals, parts/history and mobile field execution, including offline work | OSR joins ERP asset/parts/people records to offline field evidence, operational/GTFS exchange and configured operator identity adapters; national transport integrations remain deployment-specific. |
| IBM Maximo Application Suite | Commercial; benchmark | Asset/work management, inspection, reliability, monitoring and predictive maintenance | OSR connects comparable asset/work record categories to public railway definitions and engineering changes; equivalent operating scale, analytics and service support require proof. |
| Bentley OpenRail Designer | Commercial; benchmark | Specialist alignment, cant, turnouts/yards, corridor modelling and design documentation | OSR carries screening geometry into quantities, IFC, procurement and cashflow; it does not claim equivalent detailed alignment design or issued-for-construction capability. |
| OSRD | Open source; benchmark | Railway infrastructure, train simulation, timetables, capacity and short-term path requests | OSR extends the workflow into vehicle/civil references, manufacture, delivery and maintenance; breadth does not establish equivalent timetable/capacity depth. |
| Eclipse SUMO | Open source; workflow | Microscopic multimodal simulation: road, rail, public transport and pedestrians | OSR uses SUMO for independent screening and adds rail engineering, asset and delivery context; numerical agreement requires matched inputs and declared tolerances. |
| QGIS | Open source; workflow | Raster/vector GIS, spatial analysis, cartography and geospatial exchange | OSR supplies city/route revisions, demand and terrain/water screening connected to service, civil quantities and delivery outputs. |
| FreeCAD | Open source; workflow | Parametric solid modelling, assemblies, drawings and extensible engineering tools | OSR supplies controlled train/station/civil families, product interfaces, BOMs and manufacturing/evidence gates around native CAD. |
| Bonsai / IfcOpenShell | Open source; workflow | Native IFC authoring, model review, coordination and programmable BIM | OSR supplies railway asset identities, IFC4.3 packages, quantities, 4D states and executable information requirements linked to revisions. |
| ERPNext / Frappe HR | Open source; integrated | Procurement, stock, manufacturing, quality, projects, assets, finance, employees and payroll | These hold OSR's native business records; OSR adds railway context, engineering-change disposition, competence/resource checks and independent inspection/handback. |
| FUXA | Open source; integrated | Web SCADA/HMI, process displays and industrial protocol connectivity | OSR supplies versioned observations, durable history/alarms and ERP maintenance linkage; the installed OSR gateway carries no railway command authority. |
| OpenProject Community | Open source core; benchmark | Project/work-package tracking, collaboration, Gantt schedules and dependencies | OSR adds finite-resource CPM, railway quantities/BOM demand, supplier order-by dates and local/import cashflow; no OpenProject connector is claimed. |
| CalculiX / OpenSees | CalculiX GPL; OpenSees use restrictions; workflow | Finite-element analysis; structural/geotechnical nonlinear and dynamic research | OSR drives source-bound beam–pier–foundation comparisons and retains failed cases, tradeoffs and model limits; solver convergence does not qualify a railway structure. |
Why use OSR? A corridor or component change can be reviewed alongside its asset quantities, manufacturing work, procurement, cash needs and maintenance consequences, with inspectable source and retained evidence. Start with the owner/operator platform; use the competitive position and proof roadmap to judge maturity. The six-item operations implementation adds bounded resource/fleet/staff planning, offline evidence, reliability validation and operational exchange; measured/operator commissioning remains separate.
The physical reference is a 49.5 m, three-car driverless light metro with 360 AW2 / 480 AW3 capacity and 675 kWh gross / 540 kWh usable LFP storage, integrated battery/cabin thermal control, seven station archetypes and reusable civil families. A welded steel safety frame, non-structural GFRP cladding and supplier-neutral COTS subsystem interfaces support local fabrication, assembly and repair. Its 120 LM3 product rows, 26 assembly nodes, 30 tooling/mould families and 146 native FreeCAD models with matching split IFC4.3 files connect that make/buy strategy to BOMs, tooling and qualification gates.
The city catalogue uses elevated land sections in its controlled central study areas, radial tangents and curved ring connections where terrain and water permit; retained shore detours and priced bridge candidates replace long lake crossings. Outer approaches retain their actual at-grade, elevated and special-crossing mix, with no tunnels. Each city reports its current civil quantities, one full-fleet planning depot per line, graded staffing and an 18-month factory-readiness plan. These are planning concepts: property, survey, supplier, structural and operational releases remain open. Catalogue design review · Water and interchange checks · Baghdad’s dedicated proposal and funding.
The civil workbench compares complete beam–pier–foundation packages, with Baghdad as the qualification target. It searches for separate low-cost, low-mass and short-construction-time choices and retains the Pareto tradeoffs, failures, changed parameters and evidence gaps. The retained LM3 verification studies and Baghdad research results have distinct load bases; Baghdad supplier axle positions are still missing.
| Part of the system | Choices investigated |
|---|---|
| Beams and decks | Pi20/Pi25 controls; hollow box; prestressed I; U; ribbed; targeted UHPC ribs; FRP shell/concrete; segmental box; steel–concrete I; FRP–concrete I. Research spans include 20, 25 and 30 m. |
| Materials and composites | Normal, high-performance and structural lightweight concrete; targeted UHPC; steel; GFRP and CFRP orthotropic matrices. Properties remain declared research hypotheses until measured. |
| Piers | Solid; solid tapered; hollow tapered; hollow prismatic; segmental hollow; experimental double-skin FRP/concrete. Reinforcement, confinement, joint and cyclic acceptance remain open. |
| Foundations | Reference pile group; four/six bored piles; single shaft; CFA; displacement piles; hollow and square driven precast piles; spread footings. Diameter, length and cap sizes are investigated with each package. |
| Assembly and delivery | Simple spans, link slabs and continuity; full precast, segmental, shell/infill and in-situ construction; finite factory/rig/transport/erection resources, curing and inspection. |
Executed comparison: 731 distinct complete packages, 779 evaluations, 24 detailed packages, using 300 m of double track.
| Conditional objective | Beam | Pier | Foundation | Synthetic installed cost | Installed study mass | Working time |
|---|---|---|---|---|---|---|
Cheapest in concrete-first |
U-girder 20 m, lightweight | hollow-tapered | CFA-six | $1,774,796 | 6,504.4 t | 120 working days |
Lightest in concrete-first |
hybrid-shell 20 m, high-performance | hollow-prismatic | square-driven-six | $7,374,804 | 5,499.2 t | 154.625 working days |
Fastest in concrete-first |
uhpc-ribbed 20 m, high-performance | hollow-prismatic | CFA-six | $2,640,915 | 7,609.1 t | 63 working days |
These are the best confirmed research choices found within the declared search budget and synthetic price/productivity scenarios. Supplier-priced cheapest design, railway qualification and global optimality remain unresolved. All physical and operating release flags remain false.
Reduced native models screen the packages. A diverse shortlist receives three actual-section CalculiX solid meshes and three 3D OpenSees meshes, including separate tracks, asymmetric uniform loading, braking sensitivity, transverse cap bending, torsion and coupled foundation translation/rotation. Composite-slip diagnostics expose dependence on connectors. Empty shipping/lifting units remain separate from complete service mass; driven-pile production and delivery are included in construction planning. The detailed shortlist contains 16 designs at 20 m, six at 25 m and two at 30 m. All 24 converged; 22 passed every shared provisional constraint. The two 30 m cases remain rejected diagnostics, with their failed limits retained. Every refined load case repeats the shared drift, settlement, foundation-demand, deflection and lifting checks. Force and central solid-stress convergence are recorded as well as displacement convergence. Evolution considers all three commercial scenarios, whole-life cost and erection mass; per-seed benchmarks retain unique evaluations, timing, coverage and observed Pareto quality.
Native FreeCAD 1.1.4 verification passed all 25 registered beam, pier and foundation families and all 24 complete shortlisted assemblies, including valid kernel solids, independent net volumes and agreement with the retained IFC quantities. The source-bound receipt is separate from ordinary Python CAD fallback tests. Reproduce with FreeCADCmd tools/automation/verify-civil-native-cad.py in a FreeCAD environment. Structural construction-stage load paths, V/Y supports, haunched sections and calibrated ground improvement remain explicit model extensions; measured inputs, physical validation and independent acceptance remain open.
Reproduce with tools/automation/osr-python tools/automation/civil-study.py system-study and system-verify; see the retained comparison below for the complete commands. Open the generated report.html to filter choices and compare objectives, or read the retained comparison and implementation/acceptance details. The scope includes materials, reinforcement/tendon allowances, foundations, connections, production, transport, plant, temporary works, labour, QA, site carrying cost, maintenance and bearing/joint replacement scenarios. The shared engineering demonstrators reconcile native CAD mass/CG/inertia, hollow sections, IFC analysis identities, joint screens and evidence-gated mixed searches. Actual Baghdad supplier prices and construction commitments remain open. Static distributed planning loads support research comparisons; actual axle loading, calibrated soil/material/connection behaviour, physical tests and independent acceptance are required for railway qualification. No numerical winner automatically changes the catalogue, city finances or release status.
flowchart LR
P[1. PLAN<br/>City Studio / GIS] --> E[2. ENGINEER<br/>CAD / IFC4.3 / simulation] --> D[3. DELIVER<br/>CPM / ERP / QA / manufacturing] --> O[4. OPERATE<br/>assets / SCADA views / evidence] --> R[maintain / renew] --> P
The Digital Standards Thread connects publisher records and deployment law to controls, requirements, hazards, design items, evidence and G1–G4 decisions. Local validation rejects stale reviews, broken mappings and changed hashes, while its impact index identifies reassessment after change. A green result means evidence coherence—not clause conformity, physical qualification, independent assessment or legal authorization.
1. Turn city data into a costed railway candidate. Rust planning and simulation work with deterministic Python GIS/design generators to connect source-locked city data → route and station candidates → reference vehicles and civil quantities → service, fleet and energy checks → procurement, construction schedule and cashflow. In City Studio, edit lines, stations and service over 20 layers including roads, buildings, population/demand, open-DEM elevation/slope and mapped water. Compare catchment and service implications, identify likely crossings and station exclusions, seal a candidate revision and regenerate its delivery twin. This makes alternatives and their costs reviewable from the same inputs. Surveyed walking access, turnout/siding layouts, calibrated running times and engineering release require further evidence; route coverage, battery suitability and safety headways are checked against declared inputs rather than guaranteed by automation. The population, transfer and viaduct-clearance review separates native population catchments from routing demand, includes transfers through intermediate lines, and records beam/building, support/foundation and terrain conflicts. Unknown heights and incomplete mapping remain explicit release gates.
2. Engineer the railway and its interfaces. Export OSR-ALN, GIS and IFC4.3 data for QGIS and Bonsai/IfcOpenShell; generate quantities, classifications, asset identities, 4D states, IDS requirements and BCF-oriented evidence; and use native FreeCAD train, station and civil geometry. System simulation connects trains, stations, energy, wayside, points/crossings, fares, regenerative braking and depots. Route changes flow into named compatibility checks for gradient, curvature, cant, platforms, clearances, braking, energy, crosswind, recovery and evacuation. The design-detail register controls terrain/water impacts, BIM groups, vehicle datums, mechanical interfaces, load cases and route gates. All 146 split LM3 IFC files carry provenance and interface properties; the buildingSMART IDS requirements execute 4 specifications and 945 checks. This closes information delivery—not supplier selection, calculations, tolerance stacks, structural proof, physical tests or authority acceptance.
3. Deliver through local production and accountable project controls. Turn city design into asset registers, BOM demand, finite-resource schedules, supplier candidates, order-by dates, cash requirements, QA gates and construction states. ERPNext holds native projects, procurement, inventory, manufacturing, quality, finance, people and maintenance records. The first-article baseline controls 120 product rows and 26 assembly nodes; 30 tooling families, 16 factory packages and 29 drawing-definition seeds feed nested BOMs, methods, travelers and hold points. Across 62 locally made rows, 472 fail-closed slots—including exterior finish and mass closure—prevent first-article release without required production and approval evidence.
4. Operate, maintain and retain evidence. One Workbench joins City Studio, simulation, OCC, railway works, ERPNext/Frappe HR, FUXA and lifecycle views. Nine native Rust evaluator families publish exact units, identities and provenance through a versioned observation-only contract into durable history, alarms and read-only supervision. Reviewed conditions become ERP maintenance cases and Asset Repairs with parts, technicians, downtime and evidence. Local end-to-end acceptance exercises complete-city load, queued delivery, restart persistence, backup and clean-volume recovery. ERP, FUXA and management automation have no path to movement authority, point/barrier command, safety release or engineering acceptance.
- Select a catalogue city in City Studio, then inspect roads, buildings, demand, existing rail and places. 2. Draw candidate lines and stations; compare OD demand, fleet and timetable implications. 3. Add source-locked open DEM and water evidence. The router classifies likely at-grade, viaduct and bridge segments and rejects stations in mapped exclusions. 4. Model constrained single-track sections with passing at stations and recovery sidings where appropriate. 5. Compile each candidate into an immutable revision and generate its asset, quantity, cost, cashflow, critical-path and survey-gap outputs. 6. Use the comparison to commission field survey, geotechnical and hydraulic work—not to release a route or structure.
- Seal the revised GIS/OSR-ALN candidate so every downstream result refers to the same bytes. 2. Regenerate civil/station IFC4.3, quantities, classifications, asset IDs and 4D construction states. 3. Run route–vehicle gates for gradient, curvature, cant, platform, clearance, braking, energy, crosswind, evacuation and recovery interfaces. 4. Validate the 146-file LM3 federation against the IDS requirements, then run deterministic FMEA/standards and baseline-drift checks. 5. Recalculate BOM demand, finite-resource CPM, suppliers, order-by dates, QA gates and local/import cashflow; expose affected open ERP work for disposition. 6. Stop at unresolved calculations, supplier data, physical tests or approvals. A green digital report is design evidence, not construction release.
- Start from one of 120 controlled product rows and follow its CAD/IFC, mechanical interfaces, material/process fields and nested BOM. 2. Issue a bounded supplier package separating NRE, tooling, first article, repeat units and support; treat catalogue vendors as leads until qualified. 3. Create the ERP project, purchase and receipt records; preserve batch/serial identity and inspection references. 4. Execute the factory traveler and hold points, record NCRs and accepted evidence, then install the serialized item against its planned asset identity. 5. Bind design, execution, inspection and simulation commissioning evidence. A later replacement keeps the removed serial's history and requires fresh evidence. 6. Provision a lookup-only QR only after an operator has configured a secure resolver and independently verified the physical binding; see the governance templates.
- The Rust evaluator emits unit- and source-qualified observations through the command-free
osr-supervision-contractboundary. 2. The gateway retains history and alarm state; FUXA shows condition but cannot grant movement authority or engineering release. 3. A reviewed actionable condition creates one durable ERP Issue, including response guidance and immutable incident provenance. 4. A maintainer previews the linked Asset, parts availability, technician and downtime, then creates one unsubmitted Asset Repair. 5. Native ERP records work, stock and actual downtime; separate inspection and handback evidence decide whether railway use may resume. 6. Restart, ERP outage and duplicate-delivery tests prove queued events and serial history recover without turning case closure into safety clearance.
- Run normal, peak, continuously degraded and recovery profiles against the current controlled Samawah scenario using the multi-day soak harness. 2. Exercise shared battery/cabin thermal control: pack cooling, traction derate and protection take priority over comfort. 3. Remove the network path. Stored plan plus sensor localisation may retain route intent only when two fresh, trusted inputs match exactly; disagreement holds. 4. Rehearse corridor recovery from every-third-station sidings with the fail-held shunt-robot model, while retaining separation, route locking, detected points, speed supervision and emergency braking requirements. 5. Review bounded queues, historian tiers, events, CBM state, work-order evidence and the resilience report; proceed to HIL and physical trials before any safety claim.
- Freeze the ERP/SCADA context and ask the experimental executive council for an allowlisted budget, maintenance, material-request or work-order proposal. 2. Collect blind ballots from strategy, finance, operations and risk identities spanning at least two providers and three model families. 3. Require 75% endorsement; any rejection, missing perspective, stale context or diversity failure routes the proposal to human review. 4. Deterministically verify the policy, context, proposal and ballot hashes, then require an operator-keyed attestation. 5. Let ERPNext independently verify and map the result to an unsubmitted draft. Models cannot pay, contract, hire/fire, certify, release engineering, operate SCADA or command trains.
For a single runnable story joining planning, ERPNext, manufacturing, FUXA, native Rust evaluation, maintenance, restart and clean-volume recovery, follow the Samawah example-city acceptance.
| Product | Current maturity | Credible use now |
|---|---|---|
| Design and delivery platform | Serious demonstration | City GIS, alignment, service, cost, IFC, procurement, schedule, project twin and engineering coordination. |
| Train and infrastructure reference system | Engineering development | Local-manufacture planning, supplier RFQs, prototype definition, BIM coordination and civil option studies. |
| Open GoA 4 control system | R&D / pre-certification | Deterministic simulation, shadow mode, formal review and HIL preparation—not live railway command. |
The first adoptable product does not require a new autonomous railway. Start with an existing depot, workshop or pilot corridor and deploy the non-safety owner/operator stack: asset register, project controls, simulator, Ops Core, QA, maintenance and evidence portal. See the first-adoptable-product boundary.
| Unified lifecycle workspace | ERPNext city execution | Embedded supervision |
|---|---|---|
 |
 |
 |
| City Studio | Civil IFC coordination | Trainset assembly |
|---|---|---|
 |
 |
 |
These are captures of the installed simulation platform and generated engineering models, not UI mockups. Reproduce the Workbench captures with capture-platform.mjs, or watch the 88-second product assembly and 48-second civil IFC sequence.
The platform is designed to let a public owner competitively procure ordinary civil works, vehicle structures, GFRP panels, interiors, wiring, installation and maintenance locally, importing specialist components where local suppliers are not yet qualified. The reference model uses about $0.9M per 3-car light-metro trainset (current build record: $885k), a factory-gate planning target whose BOM, labour and contingency basis remains subject to supplier/process qualification. Delivered procurement must separately price engineering, homologation, commissioning, warranty and spares. The $60k per supported vehicle/car module country factory allowance scales with fleet size; national capital uses the larger of that allowance and the physical city-order factory envelope. Workshop budgets must account for the actual buildings, equipment, tooling, test paths and production-readiness work. The Iraq strategy publishes conditional city studies; a consolidated national budget, funded deployment and realised national saving remain unestablished. Transparent code and COTS choices expose assumptions for review; independently calibrated demand, risk and comparable bids are still needed to control optimism bias.
For an illustrative $100M OpenSourceRail scope, the editable default comparison applies a 2.0× foreign-turnkey price with 90% requiring foreign currency or international capital:
| Same modelled railway scope | Localisation-first OpenSourceRail | Foreign-turnkey sensitivity |
|---|---|---|
| Programme price | $100.0M | $200.0M |
| Value not requiring external capital | $79.1M | $20.0M |
| External-capital requirement | $20.9M | $180.0M |
In that scenario, the external-capital requirement is $159.1M (88.4%) lower before interest. Across the 265-city model, about $488B—roughly 79% of programme value—is assigned to domestic activity. Scope snapshot: 7 October 2026, retained catalogue finance; excludes adoption of the revised Baghdad scenario. These are reproducible planning sensitivities, not bids, audited origin claims or financing offers. Review the assumptions, low/default/high comparisons and financing cases in the portfolio calculation.
On Debian, Ubuntu, Mint, Fedora, RHEL, Rocky, AlmaLinux, CentOS, openSUSE or Arch Linux, install, build and launch the Workbench; or run the deterministic simulator and test harness:
./install.sh
./osr build
./osr book
./osr
./osr sim --duration 3600 --status-every 300
./osr testOpen http://127.0.0.1:8090/. The shared design workflow adds local jobs, validated installed BOMs, guided manufacturing/site/foundation inputs, independently reviewed revisions, broader civil/vehicle alternatives and scoped evidence verification; see the 36-package acceptance status and capability directory. The local server is not an authenticated public deployment. The installer reports existing tools before changing anything and offers larger engineering applications separately. Disposable output goes under build/ and target/release/; public review artifacts remain tracked outside those trees.
Source-locked inputs → validated candidate and deterministic generators → content-addressed Git-reviewable revision → GIS/OSR-ALN/CAD/IFC/cost/simulation/project-twin outputs → independent evidence, named approval and operational baseline.
This root README is the only human-facing front door. A hash proves which bytes were reviewed; it does not approve them. Generated city packages still require survey, calibrated demand, utility and ground data, supplier qualification, first-article manufacture and testing, competent engineering review and national authorization. The principal editable sources are:
| Concern | Source of truth | Generated/review output |
|---|---|---|
| Architecture and decisions | docs/ARCHITECTURE.md, accepted docs/rfcs/ |
Diagrams, guides and summaries |
| Rust behavior and simulation | crates/ and tests |
Applications, traces and coverage evidence |
| City synthesis and shared assumptions | design/city-generation/, lib/templates/, lib/recipes/, lib/city-batches/world-sample.toml |
City catalogue, service, finance and engineering packages |
| Iraqi funding and current publication | lib/templates/baghdad-programme-recalculation.toml, lib/templates/iraq-funding.toml, controlled procurement schedules and lib/templates/city-publication.toml |
Current Baghdad summary, latest scope and funding cases; original catalogue appraisals are labelled separately |
| Interactive city revisions | cities/workspaces/ |
Content-addressed candidates and exports |
| Mechanical/station/civil geometry | design/component-catalogue/src/osr_mech/ |
FreeCAD, IFC, BOM, traveler and image artifacts |
| Component and system practice research | engineering/best_practice/data/, linked to existing standards controls |
Reproducible handbook, catalogue coverage and open verification/backlog plans |
| ERP and supervision integration | deployment/erpnext/, deployment/supervision/, services/integration/ |
ERP workflows, FUXA views, history and maintenance evidence |
| Survey and released alignment | Accepted deployment GIS and OSR-ALN sources | QGIS, GeoPackage, corridor and IFC alignment evidence |
| Assurance and approval | docs/certification/, docs/safety-case/, engineering/assurance/formal/ |
Proof records, verification status and acceptance reports |
The local validation workflow keeps compute on your machine and GitHub for source and reviewed artifacts; GitHub Actions are disabled. Start with ./osr validate run --profile quick. The artifact policy retains useful CAD, IFC, PDF, image and animation evidence with a 50 MiB per-file ceiling.
Software is Apache 2.0. Control electronics and open physical designs use CERN-OHL-S v2. Documentation is CC-BY-SA 4.0. See LICENSE.md and LICENSES/ for the full texts.