# What's New & Changelog (/docs/start-here/changelog) Software Development Atlas evolves continuously as engineering practices, runtimes, and architectural patterns mature. This changelog documents significant content releases, new deep dives, visual demonstration upgrades, and architectural walkthroughs so readers can quickly find recent updates. ## How to stay informed [#how-to-stay-informed] The Atlas provides three built-in ways to discover new and recently revised content: 1. **Top Announcement Banner:** Highlights major milestone releases with direct links to summary notes. Dismissals persist locally in your browser. 2. **Sidebar Status Badges:** Newly authored lessons and recently updated topics display a visible badge (`NEW` or `UPDATED`) next to their titles in the navigation tree. 3. **Changelog & Freshness Records:** Every substantive page declares an explicit `lastVerified` date in its metadata. See [Content Freshness](/docs/start-here/freshness) for verification rules and review intervals. *** ## Releases & Milestones [#releases--milestones] ### September 22, 2026 · Frontend Modernization Judgment & Walkthrough [#september-22-2026--frontend-modernization-judgment--walkthrough] #### Engineering Judgment [#engineering-judgment] * [Frontend Modernization Decision Guide](/docs/engineering-judgment/decision-guides/frontend-modernization-upgrade-replace-wrap-delete): Choose whether to keep, upgrade, temporarily wrap, replace, locally rewrite, or delete a legacy capability using business criticality, compatibility, maintenance health, coupling, test evidence, reversibility, and migration surface. * [Modernizing a Seven-Year-Old React Application](/docs/engineering-judgment/architecture-walkthroughs/modernizing-seven-year-react-app): Trace a realistic modernization program from observation and stabilization through dependency cleanup, React/toolchain upgrade, Redux rationalization, vertical-slice migration, rollout, deletion of compatibility code, and relapse prevention. #### Quality & Pedagogical Synchronization [#quality--pedagogical-synchronization] * **Milestone Announcement Banner:** Updated the top release banner to reflect 54 new canonical lessons added across the Atlas since September 10. * **Language Purity Across Companion Lessons:** Standardized self-check mental-model summary tags across all Vietnamese companion lessons to ensure strict language purity. * **Decision Guide Micro-scenarios:** Synchronized real-world production micro-scenarios into structured Impact, Root cause, and Correct pattern across all engineering judgment decision guides. * **Protocol Terminology:** Aligned canonical `public client` terminology in Keycloak and identity-provider integration lessons with RFC 6749. *** ### September 22, 2026 · Frontend Modernization Migration & Safety [#september-22-2026--frontend-modernization-migration--safety] #### Frontend Engineering [#frontend-engineering] * [De-Overengineering the Frontend](/docs/frontend-engineering/de-overengineering-frontend): Remove accidental architecture by measuring abstractions through change locality, collapsing pass-through wrappers and false generalization, and preserving only boundaries that isolate real ownership, volatility, or risk. * [Incremental Frontend Migration](/docs/frontend-engineering/incremental-frontend-migration): Replace legacy behavior through vertical slices, explicit migration gates, shrinking compatibility adapters, controlled rollout, rollback, and deletion conditions instead of a big-bang rewrite. * [Safety Nets for Legacy Frontend Upgrades](/docs/frontend-engineering/legacy-frontend-safety-nets): Protect modernization seams with characterization tests, critical browser journeys, contract evidence, visual checks where relevant, production baselines, progressive rollout, diagnostics, and rollback. *** ### September 22, 2026 · Frontend Modernization Foundations [#september-22-2026--frontend-modernization-foundations] #### Frontend Engineering [#frontend-engineering-1] * [Legacy Frontend Assessment](/docs/frontend-engineering/legacy-frontend-assessment): Map runtime versions, dependency health, state ownership, critical journeys, production evidence, compatibility constraints, migration seams, and rollback boundaries before changing a legacy application. * [React Modernization](/docs/frontend-engineering/react-modernization): Separate React runtime compatibility from optional component rewrites, use warnings and Strict Mode as evidence, preserve temporary compatibility islands, and cut over incrementally. * [Dependency Archaeology](/docs/frontend-engineering/dependency-archaeology): Trace direct, transitive, peer, deprecated, prerelease, and browser-reachable dependencies so upgrade order and replacement work follow actual risk rather than package age. * [Redux Everywhere](/docs/frontend-engineering/redux-everywhere): Rationalize legacy global state by moving local, URL, remote, derived, and genuinely shared workflow state to the narrowest authoritative owner while modernizing Redux incrementally. *** ### September 21, 2026 · Frontend Authentication Architecture & Debugging [#september-21-2026--frontend-authentication-architecture--debugging] #### Engineering Judgment [#engineering-judgment-1] * [Frontend Authentication Architecture: SPA vs BFF with OIDC & Keycloak](/docs/engineering-judgment/architecture-walkthroughs/frontend-authentication-architecture): Compare direct browser OAuth with a Backend for Frontend by tracing PKCE, token ownership, protected cookie sessions, CSRF, API authorization, refresh, logout, and the evidence visible in DevTools. #### Backend Engineering [#backend-engineering] * [Auth Debugging Field Guide](/docs/backend-engineering/auth-debugging-field-guide): Debug redirect failures, CORS, 401, 403, issuer/audience/JWKS mismatches, realm/client roles, cookie sessions, refresh, and logout by locating the first broken trust boundary before changing configuration. *** ### September 21, 2026 · Identity Protocols & Keycloak [#september-21-2026--identity-protocols--keycloak] #### Backend Engineering [#backend-engineering-1] * [OAuth 2.0 & OpenID Connect](/docs/backend-engineering/oauth-and-oidc): Trace browser authentication and delegated authorization through Authorization Code + PKCE, distinguish ID/access/refresh tokens, validate issuer/audience/JWKS boundaries, and choose browser architectures without hiding client secrets in frontend code. * [Single Sign-On & Identity Federation](/docs/backend-engineering/sso-and-identity-federation): Separate IdP and application sessions, compare OIDC, SAML, and CAS trust flows, distinguish brokering from user federation, and reason about local versus global logout. * [Keycloak in Practice](/docs/backend-engineering/keycloak-in-practice): Map protocol concepts to realms, clients, public/confidential behavior, realm/client roles, client scopes, protocol mappers, identity brokering, browser login, and API authorization boundaries. *** ### September 21, 2026 · Property-Based Testing Search [#september-21-2026--property-based-testing-search] #### Testing & Quality [#testing--quality] * [Property-Based Testing](/docs/testing-quality/property-based-testing): Search broad input and action spaces with executable properties, domain-aware generators, shrinking, replayable counterexamples, stateful models, distribution checks, and bounded CI budgets without confusing sampled evidence with exhaustive proof. *** ### September 19, 2026 · Testing & Quality Evidence [#september-19-2026--testing--quality-evidence] #### Testing & Quality [#testing--quality-1] * [Test Strategy](/docs/testing-quality/test-strategy): Operate testing as a risk-driven evidence portfolio by placing unit, integration, contract, E2E, static, exploratory, performance, security, and post-deploy checks where they provide the best trade-off among speed, fidelity, reliability, diagnosability, and maintenance cost; control flakiness and protect critical user journeys without treating coverage counts or green CI as proof of production safety. * [Unit Testing](/docs/testing-quality/unit-testing): Operate fast deterministic behavioral checks through public outcomes, strong test oracles, equivalence classes, boundary values, parameterized cases, explicit state transitions, controlled clocks/randomness/IDs, order independence, refactor resilience, meaningful mock interactions, coverage limits, mutation testing, and the red-green-refactor workflow. * [Integration Testing](/docs/testing-quality/integration-testing): Operate production-relevant database, broker, cache, filesystem, and HTTP boundaries in controlled environments by running real migrations, transaction/constraint semantics, readiness-aware dependencies, isolated cleanup and parallelism, broker delivery/idempotency checks, bounded eventual-consistency polling, and diagnosable failure-path evidence without turning every test into full-system E2E. * [End-to-End Testing](/docs/testing-quality/end-to-end-testing): Reason about E2E as a deliberately small Critical User Journey portfolio: verify user-visible outcomes through the assembled system, use resilient locators and web-first waits, isolate browser/data/auth state, make third-party fidelity trade-offs explicit, treat retries as flakiness evidence, preserve traces and diagnostics, choose representative browser/device matrices, and separate pre-release E2E from post-deploy synthetic monitoring. * [Contract Testing](/docs/testing-quality/contract-testing): Reason about compatibility between independently changing consumers and providers through consumer-driven contracts, provider/schema conformance, minimal expectations, provider states, HTTP/message contracts, backward/forward compatibility, API/Protobuf evolution, versioned broker matrices, and deployment gates that check the versions actually communicating in each environment. #### Platform & Search Improvements [#platform--search-improvements] * **Command Palette & Quick Domain Suggestions:** Redesigned the search dialog into an intelligent command palette displaying core engineering domain suggestions, keyboard navigation (`↑↓` / `↵` / `ESC`), and a 3D subtle accent glow. * **Performance & Vercel Quota Protection:** Added 300ms debounce, client-side minimum query length filtering (\< 2 characters), instant visual loading feedback (\< 16ms), and 24-hour Edge CDN caching (`s-maxage=86400, stale-while-revalidate=604800`) to eliminate redundant serverless function invocations. *** ### September 18, 2026 · Cloud Infrastructure Operating Models [#september-18-2026--cloud-infrastructure-operating-models] #### Cloud & Infrastructure [#cloud--infrastructure] * [Containers](/docs/cloud-infrastructure/containers): Operate containers as replaceable host processes by separating OCI image identity from runtime instances, kernel namespace/cgroup isolation, ephemeral writable layers from durable storage, resource limits and OOM behavior, PID 1 signal handling, health/readiness, logging, networking, security hardening, and orchestration boundaries. * [Serverless Compute](/docs/cloud-infrastructure/serverless-compute): Reason about provider-managed execution environments by separating cold and warm lifecycle, environment reuse, scale-to-zero, concurrency versus instance scaling, synchronous and asynchronous invocation, retry/redelivery semantics, ephemeral state, quotas, timeouts, downstream backpressure, backlog recovery, and platform observability. * [Load Balancing](/docs/cloud-infrastructure/load-balancing): Operate traffic distribution by tracing L4/L7 frontends through routing rules and backend pools, defining stable health/readiness signals, choosing distribution and affinity policies, warming and draining targets safely, understanding zone and all-unhealthy behavior, and observing traffic skew, saturation, health transitions, and backend outcomes. * [Cloud Storage Models](/docs/cloud-infrastructure/cloud-storage-models): Reason about local, block, file, and object storage as different access/failure contracts; separate persistence from zonal/regional placement, IOPS from throughput/latency, durability from availability, replication from snapshots/backups, and storage tier pricing from recovery objectives. * [Cloud IAM](/docs/cloud-infrastructure/cloud-iam): Operate cloud authorization by tracing human and workload principals through short-lived credentials, role/service identities, principal-action-resource-context policy evaluation, trust and federation boundaries, inherited scope, explicit denies and organization guardrails, role-passing/impersonation escalation paths, revocation, audit, unused access, and break-glass recovery. * [Secrets Management](/docs/cloud-infrastructure/secrets-management): Operate authority-bearing secret material across generation, centralized storage, IAM-scoped runtime delivery, version pinning/aliases, caching, zero-downtime rotation, upstream cutover, connection-pool refresh, revocation, audit, incident response, reversible retirement, and OIDC/federation patterns that eliminate static CI credentials where possible. * [Kubernetes Fundamentals](/docs/cloud-infrastructure/kubernetes-fundamentals): Reason about Kubernetes as continuous reconciliation across API objects, controllers, scheduler, kubelet, Pods, Deployments/ReplicaSets, Services/EndpointSlices, probes, requests/limits, priority/preemption, ConfigMaps/Secrets, ServiceAccounts, PVC/PV storage, rollouts, ownership, status conditions, events, and evidence-driven debugging. * [Infrastructure as Code](/docs/cloud-infrastructure/infrastructure-as-code): Operate infrastructure changes by separating desired, recorded and actual state; reviewing plan/apply boundaries; protecting and locking remote state; reconciling drift/imports; preserving resource identity through refactors; constraining lifecycle/replacement behavior; pinning provider/module dependencies; and applying through short-lived least-privilege CI with bounded blast radius. * [Autoscaling](/docs/cloud-infrastructure/autoscaling): Reason about autoscaling as a delayed feedback-control loop by choosing capacity-proportional demand signals, separating desired from ready serving capacity, tuning warmup/tolerance/stabilization, bounding scale-in, accounting for provisioning/quota failures, queue backlog and age, downstream saturation, scheduled/predictive scaling, scale-to-zero, and interacting Kubernetes workload/node scaling loops. *** ### September 17, 2026 · Repository-Wide Pedagogical Upgrade & Distributed Systems Completion [#september-17-2026--repository-wide-pedagogical-upgrade--distributed-systems-completion] The Atlas adds **5 canonical distributed-systems deep dives** that separate replication mechanics, consensus agreement, temporary exclusive ownership, reliable event publication, and long-lived workflow compensation while keeping failover authority, committed history, quorum behavior, leases, fencing, dual-write atomicity, recovery, and reconciliation explicit. This release also delivers a sweeping pedagogical upgrade across all 65 canonical lessons: #### Pedagogical Upgrade: TL;DR 2.0 Across All 65 Lessons [#pedagogical-upgrade-tldr-20-across-all-65-lessons] * **Universal Rollout:** Upgraded all 65 canonical lessons (130 bilingual MDX files) to the **TL;DR 2.0** standard, prioritizing high-velocity learning and concrete production reality. * **Real-World Incident Hooks:** Every single lesson now begins with an iconic production outage or post-mortem (such as the 2017 AWS S3 cascading typo outage, Knight Capital's $440M deployment disaster, GitHub's 2018 network partition split-brain, Cloudflare's ReDoS lockup, and Healthcare.gov's launch failure), directly connecting conceptual models to operational stakes. * **Rule of Thumb Heuristics:** Embedded instant, high-leverage blockquotes (`> 💡 **Rule of thumb:** ...`) in every lesson header to give engineers an immediate decision compass within seconds of reading. * **Fatal Pitfall Warnings:** Made dangerous architectural antipatterns (`Fatal pitfall: ...`) explicitly visible before bad patterns reach production. * **Standardized 5-Takeaway Core:** Streamlined every lesson summary into 5 crisp, keyword-bolded takeaways optimized for efficient scanning and retention. * **React 19 & Next.js 16 Runtime Optimization:** Re-engineered announcement banner components for clean client hydration without inline script warnings. #### Distributed Systems [#distributed-systems] * [Distributed Replication](/docs/distributed-systems/distributed-replication): Reason about authoritative writers and ordered updates, asynchronous versus synchronous acknowledgment boundaries, replica lag, snapshot-based catch-up, failover fencing, multi-writer reconciliation, replica placement, and the boundary between replication and consensus. * [Consensus](/docs/distributed-systems/consensus): Recognize consensus as agreement on one committed history through quorum intersection, leader election and terms, replicated logs, explicit commit boundaries, majority/minority partition behavior, loss-of-quorum safety, and the latency cost of coordination. * [Distributed Locks](/docs/distributed-systems/distributed-locks): Reason about lease-backed ownership, renewal, unique owner tokens, fencing stale holders, safe unlock, contention control, lock granularity, and the boundary between coordination, consensus, database locks, and idempotency. * [Transactional Outbox](/docs/distributed-systems/transactional-outbox): Operate reliable event publication by committing business state and publication intent atomically, relaying committed outbox records, tolerating duplicate publication with idempotent consumers, preserving required ordering, and monitoring relay backlog and lag. * [Sagas](/docs/distributed-systems/sagas): Reason about long-lived distributed workflows as durable sequences of local transactions with semantic compensation, explicit pivot points, idempotent retry, unknown-outcome reconciliation, choreography versus orchestration, intermediate-state controls, and compensation observability. #### Cloud & Infrastructure [#cloud--infrastructure-1] * [Cloud Networking](/docs/cloud-infrastructure/cloud-networking): Operate reachability by tracing CIDR addressing, subnet placement, effective routes, gateways, NAT, ingress/egress policy, DNS, private endpoints, return paths, and connected-network boundaries without conflating network access with IAM or application authorization. * [Cloud Compute](/docs/cloud-infrastructure/cloud-compute): Reason about virtual machines as replaceable compute capacity by separating vCPU/RAM shape, images and bootstrap, lifecycle transitions, local versus persistent storage, zone/fault-domain placement, interruptible Spot capacity, right-sizing, and recovery boundaries. *** ### September 16, 2026 · Architecture, Reliability & Frontend Engineering Follow-up [#september-16-2026--architecture-reliability--frontend-engineering-follow-up] Since the September 10 milestone, the Atlas has merged **18 additional canonical deep dives** across software architecture, distributed-systems reliability, and frontend engineering. This entry reconciles the lessons published from September 11 through September 16 so the changelog matches the current `main` branch and production curriculum. #### Software Architecture [#software-architecture] * [Modular Monolith](/docs/software-architecture/modular-monolith): Design one deployable with enforceable capability boundaries, small public contracts, logical data ownership, and evidence-driven service extraction. * [Microservices](/docs/software-architecture/microservices): Reason about independently deployable business-capability services, data sovereignty, network failure semantics, compatibility, eventual consistency, and distributed-monolith risk. * [Event-Driven Architecture](/docs/software-architecture/event-driven-architecture): Model events as past facts and reason about pub/sub, durable streams, replay, at-least-once delivery, ordering, schemas, outbox patterns, lag, and coordination trade-offs. * [Domain Boundaries](/docs/software-architecture/domain-boundaries): Discover and enforce bounded contexts using business capability, language, invariants, transaction scope, ownership, context maps, and anti-corruption layers. * [Architecture Decision Records](/docs/software-architecture/architecture-decision-records): Record architecturally significant decisions with context, alternatives, evidence, explicit consequences, lifecycle status, review, and supersession. #### Distributed Systems [#distributed-systems-1] * [Partial Failure](/docs/distributed-systems/partial-failure): Treat timeout and network ambiguity as partial-failure evidence rather than proof of remote failure, with explicit recovery, idempotency, isolation, observability, and reconciliation. * [Timeouts](/docs/distributed-systems/timeouts): Define timeout scope, deadline budgets, cancellation, fan-out allocation, tuning from latency evidence, and stage-level observability. * [Retries & Backoff](/docs/distributed-systems/retries-and-backoff): Classify retryable failures and operate bounded retries with remaining-deadline budgets, exponential backoff, jitter, `Retry-After`, retry ownership, and fleet-level storm control. * [Delivery Semantics](/docs/distributed-systems/delivery-semantics): Distinguish at-most-once, at-least-once, and scoped exactly-once claims by reasoning about acknowledgement timing, crash windows, redelivery, idempotent consumers, ordering, poison-message handling, and atomic coordination between consumed progress and business effects. * [Distributed Consistency](/docs/distributed-systems/distributed-consistency): Treat consistency as an observable contract across replicas by reasoning about stale reads, read-your-writes, monotonic reads, causal order, linearizability, bounded staleness, quorum mechanisms, replica lag, and CAP behavior during partitions. #### Frontend Engineering [#frontend-engineering-2] * [CSR, SSR, and SSG](/docs/frontend-engineering/csr-ssr-ssg): Separate rendering placement and timing from freshness, personalization, cacheability, hydration, and hybrid route composition. * [Hydration](/docs/frontend-engineering/hydration): Reason about visible HTML versus interaction readiness, DOM reuse, initial-render consistency, mismatch causes, hydration cost, and debugging. * [Server and Client Components](/docs/frontend-engineering/server-and-client-components): Treat React Server and Client Components as execution and module-graph boundaries with explicit serialization, browser-visible data, composition, and bundle cost. * [Frontend State Models](/docs/frontend-engineering/frontend-state-models): Classify state by owner, lifetime, scope, persistence, and source of truth before choosing local state, URL state, shared stores, browser persistence, or remote caches. * [Frontend Data Fetching](/docs/frontend-engineering/frontend-data-fetching): Make remote observations explicit through stable request identity, intentional fetch placement, parallelism, freshness, cache ownership, race handling, mutation invalidation, and bounded retry. * [Component Boundaries](/docs/frontend-engineering/component-boundaries): Shape UI boundaries around cohesive responsibility, explicit state/data/action ownership, small data-in and intent-out contracts, composition, context trade-offs, and change locality rather than arbitrary file size. * [Frontend Bundle Performance](/docs/frontend-engineering/frontend-bundle-performance): Control browser JavaScript by tracing client dependency graphs, separating transfer from execution cost, deferring optional capabilities, tree shaking safely, auditing third parties, analyzing chunks and import chains, and enforcing regression budgets. * [Web Accessibility](/docs/frontend-engineering/web-accessibility): Operate inclusive interfaces with semantic HTML, accessible names, keyboard and focus behavior, resilient forms and errors, dynamic announcements, non-text alternatives, zoom/reflow/target checks, and layered automated/manual assistive-technology verification. *** ### September 10, 2026 · System Architecture & Engineering Foundations [#september-10-2026--system-architecture--engineering-foundations] This major milestone nearly doubles the Atlas curriculum, adding **33 new deep dives, decision guides, and architectural walkthroughs** spanning backend design, data consistency, software architecture, distributed systems resilience, security, and AI-native engineering: #### Backend Engineering [#backend-engineering-2] * [Backend Request Lifecycle](/docs/backend-engineering/backend-request-lifecycle): Trace an HTTP request end-to-end through DNS, TLS termination, reverse proxying, worker routing, database connection pools, and response streaming. * [API Design](/docs/backend-engineering/api-design): Design clear resource models, predictable error contracts, cursor-based pagination, versioning strategies, and idempotent write semantics. * [Application Caching](/docs/backend-engineering/application-caching): Deliberately trade freshness for reduced work using cache-aside, write-through, stale-while-revalidate, and stampede protection. * [Authentication & Authorization](/docs/backend-engineering/authentication-and-authorization): Separate identity verification from permission enforcement; evaluate sessions, tokens, RBAC, ABAC, and revocation boundaries. * [Backend Concurrency](/docs/backend-engineering/backend-concurrency): Preserve data invariants under overlapping requests with atomic updates, optimistic locking, and pessimistic row locks. * [Background Jobs](/docs/backend-engineering/background-jobs): Move asynchronous work out of the request path using durable job queues, state machines, worker pools, and at-least-once retry safety. * [Message Queues](/docs/backend-engineering/message-queues): Decouple producers and consumers with explicit delivery semantics, acknowledgments, dead-letter queues, and backpressure. * [Rate Limiting](/docs/backend-engineering/rate-limiting): Protect shared capacity with scoped identities, burst-aware admission policies, distributed enforcement, and explicit client retry signals. * [Idempotency](/docs/backend-engineering/idempotency): Make ambiguous retries safe with stable logical-operation keys, atomic reservation, request fingerprints, replay contracts, and duplicate-safe downstream effects. * [Service Resilience](/docs/backend-engineering/service-resilience): Contain dependency failure with end-to-end deadline budgets, bulkheads, bounded concurrency, retry budgets, circuit breakers, load shedding, and graceful degradation. #### Data Systems [#data-systems] * [Relational Data Model](/docs/data-systems/relational-data-model): Model durable business facts with candidate keys, relationship cardinality, integrity constraints, normalization, explicit source-of-truth ownership, and safe schema evolution. * [SQL Querying](/docs/data-systems/sql-querying): Operate relational queries by defining result grain, controlling join cardinality, separating row and group filters, handling `NULL`, ordering deterministically, paginating safely, and proving correctness before plan optimization. * [Database Indexes & Query Plans](/docs/data-systems/database-indexes-and-query-plans): Understand B-Tree indexing mechanics, composite index column ordering, covering indexes, and how to verify execution paths with `EXPLAIN ANALYZE`. * [Database Transactions & Isolation](/docs/data-systems/database-transactions-and-isolation): Master ACID guarantees, transaction boundaries, and isolation levels (Read Committed, Repeatable Read, Serializable) to prevent dirty reads, non-repeatable reads, and write skew. * [MVCC](/docs/data-systems/mvcc): Reason about snapshots, row-version visibility, isolation-level snapshot scope, writer conflicts, vacuum cleanup, long-running transactions, and transaction ID wraparound. * [Database Replication](/docs/data-systems/database-replication): Reason about WAL streaming, asynchronous and synchronous commit guarantees, replica lag, read-after-write routing, replication slots, hot-standby conflicts, failover, and fencing. * [Partitioning & Sharding](/docs/data-systems/data-partitioning): Reason about partition and shard keys, pruning, lifecycle boundaries, hotspots, fan-out, cross-shard invariants, global uniqueness, and resharding. * [In-Memory Data Stores](/docs/data-systems/in-memory-data-stores): Reason about working sets, source-of-truth boundaries, memory budgets, eviction, TTLs, persistence, replication, hot keys, sharding, and cache-failure contracts. * [Search Indexes](/docs/data-systems/search-indexes): Reason about inverted indexes, analyzers, text/keyword mappings, relevance, near-real-time refresh visibility, derived projections, freshness, reindexing, and alias cutovers. * [Object Storage](/docs/data-systems/object-storage): Reason about immutable object identity, strong read-after-write consistency, direct and multipart uploads, checksums, versioning, lifecycle policy, metadata ownership, and cross-system reconciliation. #### Software Architecture [#software-architecture-1] * [Coupling & Cohesion](/docs/software-architecture/coupling-and-cohesion): Keep change local by mapping change propagation, assigning business-rule ownership, increasing cohesion, reducing unnecessary code/data/temporal coupling, tightening contracts, and breaking dependency cycles. * [Modularity](/docs/software-architecture/modularity): Operate module boundaries by grouping coherent responsibilities, hiding volatile implementation knowledge, constraining dependency direction, breaking cycles, and enforcing public contracts. * [Layered Architecture](/docs/software-architecture/layered-architecture): Reason about logical layers, dependency direction, strict versus relaxed layering, layer-versus-tier boundaries, transaction ownership, mapping, bypass prevention, and the trade-off with feature-oriented structure. * [Hexagonal Architecture](/docs/software-architecture/hexagonal-architecture): Reason about ports defined by application purpose, driving and driven adapters, dependency inversion, test isolation, composition roots, and preserving failure semantics while keeping technology outside the core. * [Clean Architecture](/docs/software-architecture/clean-architecture): Reason about the Dependency Rule, policy-versus-detail boundaries, application use cases, interface adapters, inward source-code dependencies, boundary data, composition roots, and test isolation without cargo-culting the four-circle diagram. * [Monolith Architecture](/docs/software-architecture/monolith-architecture): Reason about a single deployment boundary, in-process collaboration, shared transactions and data ownership, whole-application scaling, release coupling, failure blast radius, and evidence-driven decomposition. #### Distributed Systems & Networking [#distributed-systems--networking] * [Timeouts, Retries & Backoff](/docs/distributed-systems/timeouts-retries-and-backoff): Bound downstream failure without causing retry storms using deadline propagation, exponential backoff with full jitter, and circuit breakers. * [DNS Resolution & TLS](/docs/web-platform/dns-resolution-and-tls): From recursive DNS lookups and caching records to TLS 1.3 handshakes, certificate chains, SNI, and session resumption. #### Security [#security] * [Threat Modeling & Least Privilege](/docs/security/threat-modeling-and-least-privilege): Proactively identify security boundaries using STRIDE, enforce deny-by-default authorization, and bound compromise blast radius. #### Delivery & Operations [#delivery--operations] * [Logs, Metrics & Traces](/docs/delivery-operations/logs-metrics-and-traces): Build production observability using structured logs with trace correlation, high-cardinality metrics, and distributed spans (W3C traceparent). #### AI-Native Engineering [#ai-native-engineering] * [Verification Loops & Machine-Checkable Guardrails](/docs/ai-native-engineering/verification-loops-and-machine-checkable-guardrails): Build robust agentic engineering pipelines using deterministic compilers, AST linters, type checks, and contracts rather than subjective model self-grading. #### Engineering Judgment [#engineering-judgment-2] * [Queue vs Event Stream Decision Guide](/docs/engineering-judgment/decision-guides/queue-vs-event-stream): Compare point-to-point worker queues (RabbitMQ, SQS) with partitioned log streams (Kafka, Redpanda) across consumer concurrency, replay, ordering, and operations. * [Reliable File Processing Pipeline Architecture Walkthrough](/docs/engineering-judgment/architecture-walkthroughs/reliable-file-processing-pipeline): Step through a large-scale asynchronous file upload, validation, chunking, and worker processing topology with transaction outboxes and dead-letter handling. #### Platform & Diagram Improvements [#platform--diagram-improvements] * **Responsive Diagram Sizing:** Mermaid diagrams now naturally conform to their intrinsic dimensions, center neatly within the reading column, scale text proportionally, and support controlled zooming without viewport bloat. * **Interactive Phase Walkthroughs:** Added step-by-step interactive phase walkthroughs for the payment ambiguity window and retry storms. *** ### September 09, 2026 · Foundation & Visual Demonstration System [#september-09-2026--foundation--visual-demonstration-system] * **Bilingual Core Architecture:** Released 100% paired English and Vietnamese companion lessons across all documentation routes. * **Interactive Learning Primitives:** Integrated the Event Loop Lab, Promise Resolution Lab, Async Waterfall Lab, and HTTP Request Path Explorer. * **Canonical Atlas Map & Coverage Engine:** Unified 12 engineering domains and concept registries in `content/atlas-map.json` and `content/learning-paths.json`. * **Engineering Judgment Section:** Introduced decision guides for Containers vs Serverless, CSR vs SSR vs SSG, and Monolith vs Microservices, alongside the Reliable Checkout walkthrough.