Spreadsheet Engine — System Design Academy

Complete System Design Tutorial

The design separates the latency-sensitive online path, the correctness-critical durable transition, and asynchronous projection or repair work. Each chapter below explains one decision shown in the diagrams.

1. Requirements and scope

Functional requirements

Execute the primary spreadsheet engine workflow and return an observable result.
Expose the concrete operations shown below with stable retry identities.
Allow operators to inspect, disable, repair, or replay ambiguous state.
Keep derived analytics and notifications off the correctness-critical path.

Non-functional requirements

Design for 1M cells/sheet · incremental recalc · versioned edits.
Protect the named business invariant under concurrency and partial failure.
Scale by an explicit aggregate/partition key without one global mutex.
Degrade intentionally when cache, queue, store, or regional ownership is unavailable.

Primary invariant: every object transition is legal and atomic at one aggregate boundary.

↑ top

2. Capacity assumptions

State assumptions before choosing technology. The supplied workload is 1M cells/sheet · incremental recalc · versioned edits. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

Dimension	Working estimate	Design implication
Traffic	1M cells/sheet · incremental recalc · versioned edits	Separate online compute, durable state, and asynchronous consumers.
Peak factor	5–10× average	Admission control, autoscaling headroom, and hot-key protection.
Stored state	write rate × retention × record bytes	Partition count, indexes, compaction, backup, and archival tier.
Network	peak operations × payload size	Batching, compression, edge placement, and egress budget.
Availability	99.9–99.99%	Multi-zone replicas, tested failover, and dependency error budgets.
Recovery	RPO/RTO by invariant	Ack boundary, replay retention, reconciliation, and operator runbooks.

Do not cargo-cult the numbers: show the arithmetic, name the uncertain variable, and say which production metric will replace the interview estimate.

↑ top

3. API and data model

SET_CELL(sheet, cell, formula, base_version)

GET_RANGE(sheet, range, version)

REGISTER_UDF(name, module_version)

UNDO/REDO(workbook, operation_id)

Authoritative records and indexes

cell(sheet_id, address, formula, value, version)

dependency(source_cell, dependent_cell)

calc_result(cell, input_versions, value, error)

operation(id, actor, edits, prior_version)

Command and event envelopes

# Command envelope
operation: SET_CELL(sheet, cell, formula, base_version)
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: spreadsheet-engine_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: spreadsheet-engine_123
event_type: spreadsheet_engine_committed
schema_version: 3
trace_id: tr_8ac...

The authoritative aggregate and idempotency record share a transactional or conditional-write boundary. Event streams, search indexes, caches, analytics, and dashboards are rebuildable projections.

↑ top

4. Why this architecture and what it costs

The boxes are not a shopping list. Each boundary exists because it owns a different invariant, scaling dimension, data lifecycle, security policy, or failure domain. The table makes those choices explicit and shows when the design should remain simpler.

Component	Owned responsibility	Why separate it	Rejected alternative	Cost and evolution trigger
Spreadsheet UI	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Parse edited formula, update dependency edges, and reject/report cycles.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Workbook Controller.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Edit introduces dependency cycle” to the external caller.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Spreadsheet UI boundary before extracting or scaling it.
Workbook Controller	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Mark transitive dependent frontier dirty and topologically schedule only affected cells.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Spreadsheet UI from also owning the contract and failure behavior of Cell Store.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Volatile function invalidates whole workbook” to Spreadsheet UI.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Workbook Controller boundary before extracting or scaling it.
Cell Store	Owns authoritative versioned state or immutable evidence. In this design it participates in: Evaluate pure functions with memoized input versions; sandbox UDFs under budget.	Defines the source of truth and commit boundary. It prevents Workbook Controller from also owning the contract and failure behavior of Formula Parser.	Let each service persist its own interpretation. That removes one clear commit boundary and makes retries, support, and reconciliation ambiguous. Here that simpler choice would couple “UDF loops or accesses network” to Workbook Controller.	Costs schema migration, replication, backup, and shard ownership. Split storage only when access patterns, retention, or compliance boundaries materially differ. Measure the latency, saturation, lag, or mismatch at the Cell Store boundary before extracting or scaling it.
Formula Parser	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Atomically publish calculated values for workbook version and notify collaborators.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Cell Store from also owning the contract and failure behavior of AST / Bytecode.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Concurrent edits change same formula” to Cell Store.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Formula Parser boundary before extracting or scaling it.
AST / Bytecode	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Persist operation log/snapshots for undo, merge, export, and crash recovery.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Formula Parser from also owning the contract and failure behavior of Dependency Graph.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Large range dependency explodes graph” to Formula Parser.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the AST / Bytecode boundary before extracting or scaling it.
Dependency Graph	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Parse edited formula, update dependency edges, and reject/report cycles.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents AST / Bytecode from also owning the contract and failure behavior of Cycle Detector.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Edit introduces dependency cycle” to AST / Bytecode.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Dependency Graph boundary before extracting or scaling it.
Cycle Detector	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Mark transitive dependent frontier dirty and topologically schedule only affected cells.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Dependency Graph from also owning the contract and failure behavior of Dirty Frontier.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Volatile function invalidates whole workbook” to Dependency Graph.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Cycle Detector boundary before extracting or scaling it.
Dirty Frontier	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Evaluate pure functions with memoized input versions; sandbox UDFs under budget.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Cycle Detector from also owning the contract and failure behavior of Calculation Scheduler.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “UDF loops or accesses network” to Cycle Detector.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Dirty Frontier boundary before extracting or scaling it.
Calculation Scheduler	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Atomically publish calculated values for workbook version and notify collaborators.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Dirty Frontier from also owning the contract and failure behavior of Function Registry.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Concurrent edits change same formula” to Dirty Frontier.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Calculation Scheduler boundary before extracting or scaling it.
Function Registry	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Persist operation log/snapshots for undo, merge, export, and crash recovery.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Calculation Scheduler from also owning the contract and failure behavior of Sandboxed UDF.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Large range dependency explodes graph” to Calculation Scheduler.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Function Registry boundary before extracting or scaling it.
Sandboxed UDF	Owns authoritative versioned state or immutable evidence. In this design it participates in: Parse edited formula, update dependency edges, and reject/report cycles.	Defines the source of truth and commit boundary. It prevents Function Registry from also owning the contract and failure behavior of Range Index.	Let each service persist its own interpretation. That removes one clear commit boundary and makes retries, support, and reconciliation ambiguous. Here that simpler choice would couple “Edit introduces dependency cycle” to Function Registry.	Costs schema migration, replication, backup, and shard ownership. Split storage only when access patterns, retention, or compliance boundaries materially differ. Measure the latency, saturation, lag, or mismatch at the Sandboxed UDF boundary before extracting or scaling it.
Range Index	Serves a low-latency, disposable projection with authoritative fallback. In this design it participates in: Mark transitive dependent frontier dirty and topologically schedule only affected cells.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Sandboxed UDF from also owning the contract and failure behavior of Version / Undo Log.	Read authoritative storage for every request. Simpler initially, but latency and source-of-truth load scale together and a hot key becomes a database incident. Here that simpler choice would couple “Volatile function invalidates whole workbook” to Sandboxed UDF.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1M cells/sheet · incremental recalc · versioned edits) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Range Index boundary before extracting or scaling it.
Version / Undo Log	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Evaluate pure functions with memoized input versions; sandbox UDFs under budget.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Range Index from also owning the contract and failure behavior of Collaboration Merge.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “UDF loops or accesses network” to Range Index.	Costs eventual consistency, schema evolution, replay tooling, idempotent consumers, and lag operations. Earn it when bursts or downstream failure must not block acknowledgment. Measure the latency, saturation, lag, or mismatch at the Version / Undo Log boundary before extracting or scaling it.
Collaboration Merge	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Atomically publish calculated values for workbook version and notify collaborators.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Version / Undo Log from also owning the contract and failure behavior of Persistence / Export.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Concurrent edits change same formula” to Version / Undo Log.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Collaboration Merge boundary before extracting or scaling it.
Persistence / Export	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Persist operation log/snapshots for undo, merge, export, and crash recovery.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Collaboration Merge from also owning the contract and failure behavior of operators and repair tooling.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Large range dependency explodes graph” to Collaboration Merge.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Persistence / Export boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Store formulas separately from computed values; the dependency graph decides the minimal recalculation frontier after an edit.

A single deployable with one transactional store is the right starting point for a small spreadsheet engine workload. The proposed boundary becomes justified when store formulas separately from computed values; the dependency graph decides the minimal recalculation frontier after an edit. Separating it contains the named failure “Edit introduces dependency cycle” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Topological incremental evaluation avoids recomputing the workbook, while volatile functions require explicit invalidation policy.

The design coordinates only where every object transition is legal and atomic at one aggregate boundary. Everywhere else it accepts bounded staleness or asynchronous propagation. The alternative—strong coordination for every projection—simplifies the mental model but increases tail latency and turns partitions into broad write unavailability.

Decision 03

Operational cost and evolution

Separate pure formula evaluation from I/O or user-defined functions so caching, determinism, and sandboxing remain defensible.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Evaluate pure functions with memoized input versions; sandbox UDFs under budget.” and extract another service only when measured throughput, blast radius, data lifecycle, or team ownership changes independently.

Why this design: the synchronous path contains only the decisions required to protect every object transition is legal and atomic at one aggregate boundary. Durable projections, notifications, indexing, analytics, and repair are isolated so their load or failure cannot corrupt the acknowledged operation.

Correctness boundary

Store formulas separately from computed values; the dependency graph decides the minimal recalculation frontier after an edit.

Scale boundary

Topological incremental evaluation avoids recomputing the workbook, while volatile functions require explicit invalidation policy.

Operational boundary

Separate pure formula evaluation from I/O or user-defined functions so caching, determinism, and sandboxing remain defensible.

↑ top

5. Critical flow and algorithm

Parse edited formula, update dependency edges, and reject/report cycles.

Mark transitive dependent frontier dirty and topologically schedule only affected cells.

Evaluate pure functions with memoized input versions; sandbox UDFs under budget.

Atomically publish calculated values for workbook version and notify collaborators.

Persist operation log/snapshots for undo, merge, export, and crash recovery.

Retry-safe mutation skeleton

identity = authenticate(request)
prior = idempotency.get(identity, request.key)
if prior: return prior.result

current = authoritative_store.read(partition_key)
next_state, events = domain.transition(current, request)
authoritative_store.compare_and_set(current.version, next_state)
idempotency.commit(request.key, next_state.result)
outbox.append(events)
return next_state.result

Exactly-once business effect: the network can duplicate requests. Stable identities, conditional state transitions, immutable event IDs, and reconciliation make the user-visible effect happen once.

↑ top

6. Caching, hot keys, and backpressure

Cache policy

Cache immutable or slowly changing reads with versioned keys, TTL jitter, short negative entries, and request coalescing.

Hot partition or tenant

Measure per-key and per-tenant skew. Replicate safe reads, isolate the offender, salt only when ordering permits, and enforce budgets.

Backpressure

Bound queues, buffers, connection pools, and concurrency. Shed low-priority work and scale from lag/oldest-item age rather than CPU alone.

Stampede and retry storms

Use single-flight/coalescing, jittered expiry, bounded retries, exponential backoff, and circuit breakers so one dependency failure does not multiply traffic.

Cache outage rule: never allow every miss to fall through simultaneously. Use local protection, request coalescing, circuit breakers, bounded pools, and controlled fallback.

↑ top

7. Partitioning and consistency

partition = stable_hash(primary_or_aggregate_key) → shard owner
write = conditional_update(expected_version)
read_projection = versioned and rebuildable

Strong boundary

every object transition is legal and atomic at one aggregate boundary. Use a uniqueness constraint, serializable/conditional write, expected version, or fencing token at this boundary.

Eventual boundary

Analytics, notifications, search, read models, and operational dashboards may lag when freshness is observable and replay exists.

Operation	Consistency	Reason
Authoritative mutation	Conditional / serializable	Protect uniqueness, ownership, money, inventory, order, or legal state transition.
Read after own write	Primary, session token, or cache warm	The caller must observe the result it just committed.
Ordinary read projection	Bounded stale where safe	Replicas and caches improve latency when version/freshness is visible.
Analytics and operations	Eventual + replay	Freshness is less important than isolating the online path.

Route every mutation for one aggregate to one owner/partition epoch.
Use cursors rather than offsets for changing ordered collections.
Retain tombstones until replication and repair windows close.
Move shard ownership through versioned maps and dual-read verification.

↑ top

8. Failure drills and recovery

Failure	Runtime behavior	Recovery proof
Edit introduces dependency cycle	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Volatile function invalidates whole workbook	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
UDF loops or accesses network	Resolve at source of truth with uniqueness, expected version, lease generation, or monotonic cursor.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Concurrent edits change same formula	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Large range dependency explodes graph	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.

Recovery checkpoint

# Recovery checkpoint
scenario: Edit introduces dependency cycle
operation_id: op_01J...
aggregate_key: spreadsheet-engine_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Persistence / Export

Every dependency needs a timeout, retry budget, circuit policy, and explicit fail-open, fail-closed, stale, or read-only mode. “Highly available” is incomplete until degraded behavior is named.

↑ top

9. Security, regions, and observability

Security and abuse

Authenticate at ingress and authorize the exact resource/tenant on every operation.
Encrypt transport and durable data; tokenize or isolate sensitive material.
Rate-limit by tenant, actor, resource, and network; detect enumeration and anomalous fan-out.
Redact secrets and personal data before logs or analytical storage.
Audit privileged mutations, support access, policy changes, and forced recovery.

Signals that prove correctness

User-visible success rate and p50/p95/p99 latency.
Timeouts, circuit state, saturation, and replica health by dependency.
Queue lag, oldest-item age, retries, and dead-letter growth.
Conflict, duplicate-suppression, and reconciliation mismatch counts.
Business drill: Edit introduces dependency cycle — alert before the user-visible invariant is violated.

Regional evolution

Begin multi-zone with tested failover, backups, and reconciliation.
Add global stateless ingress and regional read projections.
Keep one write owner per aggregate or partition where coordination is expensive.
Move ownership with epochs/fencing; avoid silent simultaneous writers.
Use expand/contract schemas, canaries, shadow reads, and reversible cutovers.

↑ top

10. Interviewer follow-ups

How are cycles detected?

Store formulas separately from computed values; the dependency graph decides the minimal recalculation frontier after an edit. Walk the concrete lifecycle step “Parse edited formula, update dependency edges, and reject/report cycles.” Then test it against “Edit introduces dependency cycle”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How does incremental recalculation work?

Topological incremental evaluation avoids recomputing the workbook, while volatile functions require explicit invalidation policy. Walk the concrete lifecycle step “Mark transitive dependent frontier dirty and topologically schedule only affected cells.” Then test it against “Volatile function invalidates whole workbook”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are ranges represented?

Separate pure formula evaluation from I/O or user-defined functions so caching, determinism, and sandboxing remain defensible. Walk the concrete lifecycle step “Evaluate pure functions with memoized input versions; sandbox UDFs under budget.” Then test it against “UDF loops or accesses network”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do concurrent edits merge?

Store formulas separately from computed values; the dependency graph decides the minimal recalculation frontier after an edit. Walk the concrete lifecycle step “Atomically publish calculated values for workbook version and notify collaborators.” Then test it against “Concurrent edits change same formula”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are volatile functions handled?

Topological incremental evaluation avoids recomputing the workbook, while volatile functions require explicit invalidation policy. Walk the concrete lifecycle step “Persist operation log/snapshots for undo, merge, export, and crash recovery.” Then test it against “Large range dependency explodes graph”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

Answer pattern: state the invariant, choose a mechanism, walk one failure, name the recovery signal, and explain the cost of the alternative.

↑ top

11. How to present this in 45 minutes

Time	What to cover
0–5 min	Clarify scope, workload, SLOs, and the correctness invariant.
5–10 min	Estimate throughput/storage and define concrete APIs and records.
10–20 min	Draw online, durable-write, and asynchronous/repair paths.
20–30 min	Explain the exact lifecycle, partition key, and concurrency control.
30–38 min	Discuss caching, hot keys, replication, and backpressure.
38–43 min	Run failure drills, security, observability, and regional evolution.
43–45 min	Summarize the five most important choices and trade-offs.

Final decision summary

Invariant	every object transition is legal and atomic at one aggregate boundary.
Partitioning	Stable aggregate/key hash with versioned ownership and measured hot-key isolation.
Commit	Expected version or uniqueness at the authoritative state boundary.
Async work	Durable outbox/log, idempotent consumers, checkpoints, and replay.
Recovery	Same operation identity, durable truth lookup, and business reconciliation.

Closing summary

CLOSING “The design protects every object transition is legal and atomic at one aggregate boundary. The online path stays bounded, authoritative state is versioned and partitioned, and asynchronous work is replayable and observable.”

↑ top

Spreadsheet UI

Workbook Controller

Cell Store

Formula Parser

Complete System Design Tutorial

1. Requirements and scope

Functional requirements

Non-functional requirements

2. Capacity assumptions

3. API and data model

Authoritative records and indexes

Command and event envelopes

4. Why this architecture and what it costs

Chosen boundary vs simpler monolith

Correctness vs availability

Operational cost and evolution

5. Critical flow and algorithm

Retry-safe mutation skeleton

6. Caching, hot keys, and backpressure

7. Partitioning and consistency

Strong boundary

Eventual boundary

8. Failure drills and recovery

Recovery checkpoint

9. Security, regions, and observability

Security and abuse

Signals that prove correctness

Regional evolution

10. Interviewer follow-ups

11. How to present this in 45 minutes

Final decision summary

Closing summary