Topic Backlog

Auto-maintained by research agents. Topics discovered during expansion that need coverage. Priority: HIGH (interview-critical) | MEDIUM (important) | LOW (nice-to-have) Last updated: 2026-10-17 (Round 16 recorded: research-projects expansion — 43 pages, part now 95 pages) Last content change: 2026-10-17 (43 new chapters across the six research-project collections)

Round 16 — Research projects expansion — 2026-10-17

• User-directed expansion of the six collections: 43 new pages (follow-up tree 14 chains total, assumptions hunt 14 audits, connection atlas 16 connections, why-not handbook 19 comparisons in 14 files, failure casebook 12 cases, misconceptions 18 claims). SUMMARY.md part entries extended to 95; per-project Coverage Ledgers updated; hub index counts refreshed.
• New primary-sourced case histories: Ariane 5 (Lions report), Heartbleed (OpenSSL advisory), GitLab 2017 (postmortem), CrowdStrike 2024 (RCA PDF + mirrors), Fastly 2021 (official summary via contemporaneous verbatim excerpts), AWS us-east-1 2021 (AWS message).
• Candidate follow-ups: remaining domains for follow-up chains (consensus, CRDTs, TLS 1.3 handshake); CRDT-vs-OT and capability-based security for the atlas; additional verified incidents for the casebook — MEDIUM; all require source verification before writing.

Round 15 — Six research-project collections — 2026-10-17

• New top-level SUMMARY part src/research-projects/ (52 pages, ~109K words, 10 Mermaid diagrams): follow-up tree (6 escalation chains), hidden assumptions hunt (7 audits), concept connection atlas (9 connections), why-not handbook (12 comparisons in 7 chapters), real-world failure casebook (6 primary-source incident case histories), misconception elimination (10 audited claims).
• Coverage ledgers per project live in each index.md; hub at index.md. Cross- linked to existing chapters (trade-offs, failure-modes, sre, interview, dsa, os, dbms, networks, concurrency, distributed, references).
• Candidate follow-ups spotted while writing:
  • More follow-up chains for high-demand domains (caching layers, Kafka, Kubernetes scheduling) — MEDIUM.
  • Additional incident case histories with verifiable primary postmortems (e.g., Node.js/npm ecosystem incidents, Kubernetes CVE-linked outages) — MEDIUM; each requires source verification before writing.
  • Agent 3-a noted datatracker.ietf.org returned 403 from the work sandbox; RFC links were kept on the strength of the RFCs being canonical, but a later pass may re-verify from a different environment — LOW.

Round 14 — Section G eight-language process/runtime + navigation integrity — 2026-10-13

• User-directed round: a full exit-codes/signals/process-lifecycle knowledge base for 8 languages (C, C++, Python, Go, Java, Rust, Node.js, OCaml) landed as src/languages/process-runtime/ (18 pages: 8 chapters + 8 ledgers + hub + 9-dimension comparison), audited against the G1–G6 acceptance criteria with 633 ledger rows (623 documented / 10 source-limited). A G3 reference-policy pass added 144 verified hyperlinks where chapters had cited sources by name.
• Maintenance round in the same window: 295 orphaned pages wired into SUMMARY.md (4,116 entries now, 0 dup / 0 dead, mdBook v0.5.4 build clean) and 130 mdBook “unclosed HTML tag” warnings cleared across 33 files.
• Candidate follow-ups spotted while auditing:
  • 217 pre-existing broken internal links/anchors repo-wide (scripts/check-links.py; stash-baseline-proven pre-existing, none introduced by recent rounds) — the next HIGH-value maintenance pass; each needs a re-point / retarget / drop decision.
  • 10 Section G ledger rows stay source-limited (glibc manual and en.cppreference.com unreachable from the work environment) — retry from an environment with access.
  • Carried from round 11: stale section-README tables (security/advanced lists 5 of 22 files; several mindmaps predate the new pages — SUMMARY remains the source of truth).

Round 13 — Interface Encyclopedia per-function rebuild + process-exit-signal KB — 2026-10-12

• User-directed round replacing mega-chapters with one page per interface: syscalls 207 files, libc 292 (glibc families + musl catalog), sysfs 121 (ledger + 119 classes), crossref 5 — with source-regenerated ledgers and statuses flipped honestly (101 syscall rows, 10 glibc rows, 14 sysfs classes, 1 crossref row stay source-limited with named reasons; 46 syscall rows are researched and upgradeable).
• New part src/linux/process-exit-signal/ (43 pages): signals inventory, exit-status lifecycle, shell chapters (bash/dash/zsh), service managers (systemd/OpenRC/runit/s6), containers (PID 1/Docker/K8s), troubleshooting decision tree; unified ledger of 174 requirement IDs.
• Candidate follow-ups spotted while writing: upgrade the 46 researched syscall rows to documented (sources exist, effort-bound); per-signal deep pages for the job-control signals (SIGTSTP/SIGTTIN/SIGTTOU are ledger rows but thin in prose); a compression-tool family page set (gzip/xz/zstd — carried over from round 12’s candidate list, still homeless); an ldd/ld.so/ELF-loader page set (referenced by five+ binary pages).

Round 12 — Linux userland binaries sweep — 2026-10-09

• User-directed round replacing the usual gap-mine: one page per binary for every binary in the core userland packages (collection names taken from uutils.org). 280 new pages under src/linux/binaries/ in 16 collections, each collection with a hub page carrying package-level context. Wired as a new top-level SUMMARY part “Linux Userland Binaries”.
• Dedup: find/xargs/grep/sed/awk kept their existing deep shell-track pages; hubs cross-link instead of duplicating. bsdutils binaries (logger, wall, write, script, scriptlive, scriptreplay, renice) get their pages in the bsdutils collection with cross-links to util-linux pages; login/sulogin/ nologin live in the login collection. raw(8) and pg(1) are documented as removed upstream rather than given pages.
• Reference discipline held: 285 per-binary manpage URLs verified live on manpages.debian.org (the Debian util-linux split was mapped by scraping package indexes: fdisk, mount, uuid-runtime, eject, ncal, bsdextrautils, util-linux-extra, sysvinit-utils, uidmap), plus sources.debian.org, GitHub mirrors, BusyBox manual, POSIX spec. gnu.org unreachable from the work environment → not cited except the pre-verified tar manual.
• Candidate follow-ups spotted while writing: a gzip/gzip-family page set (the compression tools are referenced everywhere but have no home — coreutils-adjacent gap), an ldd/ld.so/ELF loader page set (referenced by five+ binary pages), a sysstat (iostat/sar) collection to complete the observability story procps starts, and deepening the login collection’s shorter pages (lastlog/newgrp/sg/sulogin run 217-240 lines).

Round 11 — backlog re-mine + inventory scan — 2026-10-09

• Of the six round-10 candidates, four were already covered and dropped after dedup: gVisor/Kata/Firecracker (cloud/virtualization/ has dedicated pages), feature-store & ML-data tooling (ml/system-design/feature-store.md + the interview case study), the WASM component model (compilers/advanced/wasm-component-model.md), and eBPF observability tooling (linux/observability/ebpf-observability.md + continuous-profiling.md both carry Pixie/Parca). The other two became pages with complementary angles: database migration tooling (versioned/declarative authoring + expand– contract, NOT the pt-osc/gh-ost mechanics online-schema-change.md already owns) and LLM eval tooling (promptfoo/Inspect AI/DeepEval/Ragas/LangSmith mechanics, NOT the benchmark-harness territory eval-harnesses.md owns).
• Five new pages: Zanzibar & relationship-based access control (security/advanced/zanzibar-rebac.md), schema migration tooling (dbms/advanced/schema-migrations.md), LLM eval tooling (llm/evals/eval-tooling.md), geospatial indexing (dbms/advanced/geospatial-indexing.md), micro-frontends (frontend/micro-frontends.md).
• Two thin pages deepened with all original content preserved: sre/chaos-engineering.md 76 -> 271 and cloud/disaster-recovery.md 119 -> 260 (slo-sli-sla.md was considered but rejected — slo-error-budget.md already owns burn-rate alerting, so a deep-dive would duplicate).
• Round-12 candidates spotted while scanning: a data-contracts page (5 passing mentions; complements data-engineering/data-quality.md), email deliverability infrastructure beyond the email case study (SPF/DKIM/DMARC mechanics, bounce handling, warm-up — currently 6 passing mentions), and a housekeeping pass on stale section-README tables (security/advanced/README reading-order lists 5 of 22 files; several section mindmaps predate the new pages — SUMMARY is the maintained source of truth).

Round 10 — backlog-queue mining — 2026-10-09

• All six round-9 spot candidates became pages: Valkey & the Redis fork landscape (dbms/caching), Well-Architected frameworks (cloud), Talos/ k3s & minimal Kubernetes (cloud), offensive web tooling (security/advanced), vulnerability reachability tooling (security/advanced), and the 44-line distributed-projects stub became a 271-line lab curriculum (projects/build-it-yourself).
• Dedup notes: jepsen-maelstrom.md already covers the Glomers challenge mechanics (the new labs page is the curriculum view and cross-links it); web-security.md holds all vulnerability theory (the offensive page is tooling-only); appsec-toolchain’s reachability section links forward to the new tooling page.
• Round-11 candidates spotted while writing: gVisor/Kata/gVisor-vs-Firecracker sandboxing spectrum beyond the existing containerd/Kata pages (check cloud/virtualization first), database migration tooling deep-dive (gh-ost/pt-online-schema-change/flyway-liquibase), feature-store & ML-data tooling (feast), webassembly component model deep-dive (wasm already has a page — component model may not), LLM eval tooling beyond evals/README (promptfoo/inspect-ai), eBPF-based observability tooling deep-dive (pixie/parca — check observability dir first).

Round 9 — full-corpus mining, all 13 indexes — 2026-10-08

• Re-ran scripts/ref_gap_analysis.py over all thirteen indexes: 1,781 entries, 1,066 raw misses; heavy education/venue noise filtered, then each candidate dedup-checked against the repo (many “misses” were already covered — Go/Java/Rust concurrency, wasm, MCP, event-driven, block layer, backup). True gaps curated to 15 new pages.
• OS kernel families (5): Windows NT internals, XNU/Darwin, RTOS landscape (Zephyr/FreeRTOS/NuttX + Tock/Hubris), illumos/DTrace heritage, POSIX & SUS standards — src/os/advanced/.
• Security, first mining of the new index (4): AppSec toolchain (SAST/SCA/secrets/DAST), vulnerability management (CVE-to-patch, CVSS/EPSS/KEV), reverse engineering & malware analysis, host detection & response (osquery/Wazuh/Tetragon) — src/security/advanced/.
• DBMS (3): search-engine internals (Lucene/ES/OpenSearch), Vitess sharding middleware, query-engine frameworks (Calcite/DataFusion).
• Runtimes (3): managed runtimes beyond the JVM (PyPy/GraalVM/LuaJIT/.NET), rustc query-based internals, JAX & functional ML.
• Enhancements: cloud/autoscaling.md 148 -> 290 (HPA/KEDA/Karpenter), os/advanced/kernel-architectures.md 147 -> 251 (real-kernel survey, verification spectrum, links to the five new kernel-family pages).
• Remaining round-10 candidates spotted while deduping: Valkey, well-architected frameworks, Talos/k3s, offensive web tooling deep dive, gossip-glomers-style labs, vuln reachability tooling.

Skipped gaps + new reference indexes — 2026-10-08

• Track 1: the five deliberately-skipped reference-mining candidates got dedicated pages (HotSpot internals, Fuchsia/Zircon, Nomad, in-memory data grids, IDP/Backstage) — the “remaining known gaps” list from the mining round below is now fully closed.
• Track 2: references section grew 10 -> 13 indexes (security 81, concurrency 68, storage 64 entries = +213 entries, +522 URLs). All 321 CSV URLs HTTP-verified with redirects/25s-timeout; bot-blocked flagged in-page per the README convention; 8 dead/moved URLs replaced.
• Verification tooling: scripts/verify_urls.py in the workspace (parallel curl, records final status + resolved URL; CSV and –md modes).
• No known remaining gaps from the reference indexes; future index candidates if ever wanted: graphics/game-dev, bioinformatics, quantum.

Reference-library mining round — 2026-10-08

• Systematic gap analysis: all 888 entries of the ten src/references/ indexes diffed against the book’s page inventory (script: scripts/ref_gap_analysis.py in the workspace; ~945 raw misses after removing education-track and paper-venue noise, curated to 15).
• New pages grounded in the indexes’ verified primary sources: OS — xv6 teaching kernel, BSD family internals, kernel source-reading curriculum. Architecture — architectural simulation (gem5/Spike/ Verilator/Ramulator/DRAMsim3/CACTI), µarch measurement toolbox (uops.info/llvm-mca/Agner Fog), hardware root of trust (OpenTitan/TPM/ OpenSBI). Data — DuckDB internals, Apache Beam, time-series databases (InfluxDB/TimescaleDB, Gorilla + delta-of-delta worked examples). Distributed — Jepsen/Maelstrom/Gossip Glomers, TigerBeetle internals. Tooling — reading small compilers (chibicc/QBE/c4), IaC beyond Terraform, policy as code (OPA/Kyverno/Falco), agent SDK landscape 2026.
• Remaining known gaps (deliberately skipped — passing mentions already exist or marginal interview value): HotSpot-specific internals page (JVM covered via jvm-jit/gc/bytecode), Fuchsia/Zircon dedicated page (kernel-architectures.md covers microkernel theory), Nomad deep dive, in-memory data grids (Hazelcast), Backstage/IDP platform page.
• All pages wired into SUMMARY; validators green; real Mermaid parser 5,447/5,447.

Reference-grounded expansion — 2026-10-07

• 140 new pages across 15 new sections (LLM architectures / post-training / agentic / prompting / retrieval, system-design case studies, coordination & messaging internals, lakehouse formats, database internals, modern kernel, OS security internals, network protocols, cs-theory, DSA ch195-202).
• 10 previously-thin README/overview pages deepened; cs-theory thin pages expanded (logic, proofs, formal languages, computability, Turing machines, complexity classes, comparison-sorting lower bound).
• Backlog items closed this pass: Ceph CRUSH/RADOS deep dive.
• All pages wired into SUMMARY; validators green; real Mermaid parser 5,182/5,182.

Competitive/math/quant round — 2026-10-08

• Three new top-level sections closed the competition-track gap: competitive programming (8 pages — CF/AtCoder/LC guides, ICPC/IOI, clist.by/vJudge calendars, judge directory), competitive mathematics (8 pages — IMO, olympiad NT/comb/algebra/geometry, India IOQM→INMO, resource directory), and quant firm puzzles (9 pages — firm directory, Jane Street/HRT puzzle cultures, EV derivations, market-making games, mental math, game theory, canon map).
• Cross-wired with existing content: dsa game-theory chapters (ch61/162/ 180/198), mathematics/, interview/puzzles/, aptitude/.
• All 27 pages wired into SUMMARY; validators green; real Mermaid parser 5,407/5,407.

Placement-process & puzzles round — 2026-10-08

• New puzzles section (4 pages, worked solutions) — closed the brain-teaser gap for interview warm-ups.
• All 11 placement-preparation pages deepened (funnel, campus mechanics, technical/HR/GD/communication, OA/cognitive/coding assessments, internships/PPO) + aptitude hub/trigonometry/calculus expanded.
• Validators green; real Mermaid parser 5,354/5,354.

Interview-classics expansion — 2026-10-08

• 36 new pages across interview case studies (18), machine coding (9), security (3), modern kernel (3), HPC (2), and networks DDoS (1).
• 10 thin high-traffic pages substantially enhanced (coding patterns, OA/MCQ strategies, section README catalogs, TGI, LLM serving security, vector DB internals, CI/CD case study).
• Backlog items closed this pass: IRCTC/Tatkal booking, UPI/NPCI, video conferencing (SFU), CDN service design, DDoS mitigation, threat modeling, zero-trust, incident response, Rust-in-kernel, THP/khugepaged, zram swap, SIMD/roofline, HPC checkpoint/restart, machine-coding classics (car rental, thread pool, pub-sub, cron parser, Stack Overflow, cricket scoreboard, marketplace, content moderation).
• All pages wired into SUMMARY; validators green; real Mermaid parser 5,336/5,336.

Dev pull audit — 2026-08-13

• Pulled the latest dev expansion and repaired ten missing Summary targets plus four stale relative links. All pulled pages are now Summary-reachable.
• Remaining work is content depth/references review, not broken navigation.

Software Engineering pages — 2026-08-13

• Added dedicated testing.md, devops.md, and CONTRIBUTING.md under src/software-engineering/ (previously those links were redirected to the general testing/ and cloud/cicd/ sections and a GitHub URL).

Integration completed — 2026-08-12

• Linux deep dive (lb2) — integrated 444 educational pages under src/linux/, with source navigation adapted to the parent Summary.
• Linux tools — added a focused guide for find, xargs, text filters, process/system-call tools, storage inspection, networking, HTTP debugging, SSH/rsync, jq, make, and tmux, including interview questions, safety notes, and official references.
• DSA track (dsa_book_2) — integrated 193 educational pages and appendices under src/dsa/, from foundations to advanced algorithms.
• Integrity gate — 0 broken links, Summary navigation OK, and 4,387/4,387 Mermaid diagrams pass both the heuristic and Mermaid v11 parser checks.

Completed (since 2026-08-05) — Updated 2026-08-13

The following previously-backlogged topics now have dedicated coverage:

• Languages: C (8 files incl. ecosystem), C++ (8 incl. ecosystem), Rust (9 + ecosystem), Java (incl. GC java/gc.md + ecosystem), Python (9 + ecosystem), Go (6 incl. scheduler 5 files → 6 with web-frameworks), JavaScript/Node/V8 (3 → expanded nodejs 414 lines, v8 363 lines), OCaml (incl. ecosystem/tooling), TypeScript (type system, generics), Go Web Frameworks (NEW 2026-08-09: Gin/Echo/Fiber comparison with prod benchmarks, allocation analysis, decision flowchart)
• Frameworks: Spring Boot, FastAPI, Django, Express.js, React, Next.js, Tokio, PyTorch, Vue & Angular (combined page exists, could split)
• Backend: REST, gRPC, GraphQL, JWT/OAuth/sessions, Docker, Kubernetes, CI/CD (GitHub Actions, GitOps), Redis, Kafka, RabbitMQ, NATS, service mesh (service-mesh.md + NEW xds-protocol.md 2026-08-08: ADS, LDS/RDS/CDS/EDS/SDS, protobuf not YAML, SotW vs Delta), event sourcing, CQRS, idempotency, webhooks, connection pools, testing, xDS Protocol (NEW)
• ML/LLM: Transformers, attention, GPT/BERT, fine-tuning, RLHF/DPO, inference optimization (KV cache, quantization, speculative decoding, batching), RAG, vector databases, LLM security (OWASP Top 10), distributed training (DDP/FSDP/ZeRO/TP/PP), diffusion, MoE, Probabilistic Data Structures (NEW 2026-08-09: Bloom 10 bits/elem @1%, HLL 12KB fixed ~2%, CMS overestimates, use cases Cassandra/Chrome/BigQuery)
• OS: eBPF, io_uring, Linux kernel internals, namespaces, cgroups v2, kernel modules (NEW 2026-08-08: LKM lifecycle, insmod/modprobe, params, sysfs, signing/taint), kernel tracing (NEW 2026-08-09: ftrace, kprobes/kretprobes 0.05us vs 0.5us, tracepoints, perf, eBPF bpftrace one-liners), RCU (NEW 2026-08-09: grace period, publish-subscribe, quiescent states, checklist)
• Networks: QUIC, HTTP/3 (http3.md), QPACK RFC 9204 (NEW 2026-08-08: static 99 vs dynamic table, encoder/decoder streams, RIC/Base, blocking), TLS deep dive (1.3 handshake, ECDHE, mTLS, CT), Bluetooth (NEW: PAN profiles), eBPF networking
• DBMS: query optimization (cost model, join algorithms), storage engines (WAL, LSM), distributed SQL (CockroachDB/TiDB/Spanner), OLTP vs OLAP, WAL (NEW 2026-08-09: LSN/pageLSN, ARIES 3 phases, CLR, group commit, fuzzy checkpoint), LSM Compaction (NEW 2026-08-09: leveled vs tiered vs hybrid, WA 10-30× leveled, RocksDB tuning knobs, SAG), Congestion Control Overview (NEW: congestion-control/README.md wired)
• Cloud: autoscaling, disaster recovery & multi-region, IAM & secrets management, Kubernetes operators, Kubernetes operators (already done)
• Arch: CUDA programming (parallelism/cuda.md), Modern Processors expanded (x86-64 Xeon/EPYC, ARM Neoverse, RISC-V) via dev merge
• Interview: streaming analytics pipeline system design (real-world/streaming-pipeline.md), Probabilistic Data Structures for System Design (NEW), real-world examples Netflix/Twitter/Uber/WhatsApp/YouTube/Instagram/Dropbox/Google Search/distributed lock
• Storage: HDD/SSD/NVMe/object/block/file/distributed/Ceph/erasure-coding (10 files) → 12 files with WAL + LSM Compaction + Bluetooth fix, still low coverage (45%)
• Concurrency: fork-join/thread-pools/producer-consumer/readers-writers/async-await/futures/coroutines/lock-free/transactional-memory/go-channels/rust-ownership/java/python-gil (14) → 15 with RCU, still low (48%)
• OOP: coupling, cohesion & design principles (lld/coupling-cohesion-principles.md)
• Security: TLS deep dive, mTLS, certificate transparency already; auth expanded (OAuth, JWT, session, mTLS)

NEWLY COMPLETED (2026-08-13 — Parallel Agent Batch)

Completely New Sections Added:

• Git ✅ src/git/ — 15 files: internals, fundamentals, branching, rebasing, stashing, advanced ops, remotes, tags, hooks, workflows, GitHub, interview questions, cheat sheet
• Software Engineering ✅ src/software-engineering/ — 10 files: SDLC, Agile, requirements, design (SOLID), code quality, documentation, project management, metrics, team dynamics
• Programming Fundamentals ✅ src/programming-fundamentals/ — 10 files: variables/types, scope/lifetime, functions, error handling, type systems, memory model, I/O, modules
• Security & Cryptography ✅ src/security/ — 7 files: authentication (OAuth/OIDC/JWT), authorization (RBAC/ABAC), web security (OWASP), cryptography (AES/RSA/TLS), secrets management
• Machine Coding ✅ src/machine-coding/ — 10 files: parking lot, elevator, library, splitwise, rate limiter, LRU cache, task scheduler
• Data Engineering ✅ src/data-engineering/ — 7 files: Spark, Kafka, Airflow, Parquet/Avro/ORC, data quality
• Search Engines ✅ src/search/ — 5 files: inverted indexes, TF-IDF/BM25, Elasticsearch, vector search
• Aptitude ✅ src/aptitude/ — 11 files: percentages, ratios, averages, profit/loss, time/work, speed/distance, probability, number systems, logical reasoning, data interpretation
• Placement Preparation ✅ src/placement-preparation/ — 6 files: campus process, OA, technical/HR interview, group discussion
• Resume & Career ✅ src/resume/ — 7 files: structure, bullet writing, projects, ATS optimization
• Behavioral Interviews ✅ src/behavioral-interviews/ — 5 files: STAR method, 30+ questions, company fit, scenarios
• Communication ✅ src/communication/ — 4 files: technical, interview, written communication
• Practical Programming Problems ✅ src/practical-problems/ — 6 files: parsers, CLI tools, file processing, concurrent problems
• DBMS Interview Problems ✅ src/dbms/interview-problems/ — 6 files: classic SQL, window functions, joins, optimization, concurrency

REMAINING — MEDIUM Priority (Updated)

Operating Systems — Partially Done

• Linux Kernel Modules — loading, unloading, parameters (covered briefly; expand into a page) ✅ Done os/kernel/modules.md
• Linux Tracing Deep Dive — ftrace, kprobes, perf, eBPF, and advanced tracing are covered in the integrated Linux track; an LTTng comparison remains optional
• Memory Barriers Deep Dive — C++20/Java/Go memory models (acquire/release, seq_cst) — dedicated page missing, currently in OS sync

Networks — Partially Done

• HTTP/3 QPACK — header compression specifics ✅ Done networks/http/qpack.md
• Service Mesh internals — Istio/Envoy xDS protocol, sidecar dataplane ✅ Done backend/containers/xds-protocol.md
• eBPF Networking Deep Dive ✅ Done networks/ebpf-networking.md (2026-08-12): XDP, TC, socket hooks, maps, AF_XDP, Cilium, CO-RE, and observability

Languages / Frameworks — Partially Done

• Go web frameworks — Gin/Echo/Fiber comparison ✅ Done languages/go/web-frameworks.md
• Rust Async Runtimes ✅ Done languages/rust/async-runtimes.md (2026-08-12): Tokio, smol, async-std, Glommio, Monoio, Embassy, blocking and cancellation
• Python Free-Threaded (3.13t) Deep Dive — GIL removal, free-threaded build
• Java Loom Virtual Threads — expand java.md with Project Loom

Backend / Storage / Concurrency — Newly Discovered

• Tiered Storage and Data Temperature ✅ Done storage/tiered-storage.md (2026-08-12): hot/warm/cold, RocksDB, object lifecycle, caches, recovery
• Storage: SSTable Format — data blocks, index, bloom, footer, compression ✅ Done storage/sstable.md (2026-08-09): BlockBasedTable diagram, Index/Bloom/Footer 48 bytes magic, partitioned index/filter, read path, compression
• Storage: BlobDB — separation of small vs large values ✅ Done storage/blobdb.md (2026-08-09): WiscKey, BlobIndex file_no/offset/size, GC age cutoff 0.25, WA 1.4-1.7 vs 6.1-6.8 75% lower, options enable_blob_files/min_blob_size
• Storage: Ceph CRUSH/RADOS Deep Dive ✅ Done storage/formats/ceph-crush.md (2026-10-07): CRUSH map hierarchy, PG count math, chooseleaf rules, tunables, RADOS write path, recovery/backfill
• Storage: NVMe over Fabrics ✅ Done storage/nvmeof.md (2026-08-12): TCP/RDMA, discovery, queues, multipathing, security, observability
• Concurrency: Work-Stealing Scheduler — Go scheduler work-stealing, Java ForkJoinPool, Rust Tokio ✅ Done concurrency/work-stealing.md (2026-08-09): LIFO owner head vs FIFO thief tail, Go GMP P local 256 + global, Java ForkJoinPool WorkQueue 4096, Tokio 256 ring + injection queue
• Concurrency: ABA Problem & Memory Reclamation ✅ Done concurrency/aba-problem.md (2026-08-12): tagged pointers, hazard pointers, EBR, RCU, reference counting, memory ordering, C++26 safe-reclamation references
• Concurrency: Memory Model — C++/Java/Go memory models, acquire/release, data races ✅ Done concurrency/memory-model.md (2026-08-09): TSO vs weak ARM, DRF-SC happens-before, C++11 6 orders, Java volatile total order vs VarHandle acq/rel, Go channel happens-before, Rust Send/Sync, store buffering litmus

Interview / System Design — Newly Discovered

• GraphQL Federation ✅ Done backend/api/graphql-federation.md (2026-08-12): entities, composition, directives, query planning, governance
• API Versioning Strategies — URL vs header vs content negotiation, Stripe example ✅ Done backend/api/versioning.md (2026-08-09): 6 strategies table, URL path safest default, date-based Stripe pinning compatibility layer, decision flowchart, Deprecation/Sunset headers
• Rate Limiting Algorithms Deep Dive — token bucket vs leaky bucket vs sliding window logs vs sliding window counter ✅ Done backend/api/rate-limiting.md (2026-08-09): 5 algos fixed vs sliding log O(n) exact vs sliding counter 2 keys ~1% err vs token bucket HASH burst vs leaky queue steady, comparison table, Redis Lua token bucket script, decision tree
• Distributed Lock Deep Dive ✅ Done distributed/fundamentals/distributed-locks.md (2026-08-12): leases, fencing tokens, Redis/Redlock, ZooKeeper, etcd, and alternatives
• Change Data Capture & Outbox Pattern ✅ Done backend/patterns/cdc-outbox.md (2026-08-12): dual-write failure, Debezium, WAL retention, idempotency, ordering

LOW Priority

• Quantum Computing — qubits, gates, algorithms (curiosity, not interview-critical)
• Blockchain — consensus PoW/PoS, smart contracts, Merkle trees (could be arch/blockchain.md or separate)
• Edge Computing — CDN compute, edge functions (covered in cdn/edge.md; could expand with Cloudflare Workers, Fastly Compute@Edge)
• Observability Deep Dive ✅ Done backend/observability/opentelemetry.md (2026-08-12): signals, propagation, Collector, semantic conventions, sampling, cardinality

NEWLY COMPLETED (2026-08-15 — OpenClaw expansion batch)

• Master topic index ✅ src/index.md — added the 1 660-line master topic database that was previously missing from the repo, covering all 50 sections from Algorithms through Interview Meta Topics. Wired into SUMMARY.md at the top navigation and the Meta section.
• Java Virtual Threads deep dive ✅ src/languages/java/virtual-threads.md expanded from 139 → 261 lines. Added JEP timeline (JDK 19 through 25), continuation- on-heap internals, three creation patterns, structured-concurrency API churn note (JEP 499 / 505), scoped-values rationale, migration playbook, comparison table vs goroutines / Kotlin coroutines / Reactor, 8 interview questions.
• Python Free-Threaded deep dive ✅ src/languages/python/free-threaded.md expanded from 148 → 235 lines. Added PEP 779, biased-locking internals, immortal objects (PEP 683), per-object critical sections (Py_BEGIN_CRITICAL_SECTION), Cython freethreading=True directive, C-extension compatibility table, runtime introspection (sys._is_gil_enabled, -X gil=0), comparison vs Java Loom / Go / Ruby Ractor, 7 interview questions.
• Ceph CRUSH/RADOS deep dive ✅ src/storage/ceph-crush.md expanded from 181 → 340 lines. Added straw2 bucket algorithm explanation, CRUSH rule step semantics, upmap balancer override, PG peering state machine (Mermaid state diagram), Bluestore internals (WAL/BlockDB/blob sharing), replication vs erasure coding table, upper-layer services (RBD/CephFS/RGW), performance characteristics, comparison vs HDFS/MinIO/S3, 9 interview questions, references to Weil SC 2006 paper and doctoral dissertation.
• Formal Methods ✅ src/cs-theory/formal-methods.md — new 282-line page covering index.md Section 36. TLA+ (with Counter example, AWS DynamoDB case study), Alloy (with file-system example), Coq (CompCert, Four Colour Theorem), Lean (Mathlib, Verdi), Isabelle/HOL (seL4), Dafny (binary search auto-active verification), model checking (BDD, BMC, partial-order, symmetry), symbolic execution (KLEE/angr/SAGE), abstract interpretation (Astrée, Rust borrow checker), distributed-systems verification workflow, trade-offs table, 8 interview questions.

Decisions / Notes

• Keep autonomous loop adding topics based on coverage lowest first: Storage (45%) and Concurrency (48%) still lowest → prioritize WAL, LSM compaction (done), next SSTable, BlobDB, RCU (done), memory barriers, work-stealing.
• After storage/concurrency reaches 60%, move to Frameworks (60%) and Distributed (55%) and Cloud (60%)
• Update knowledge graph each batch to keep cross-links
• Maintain 100% mermaid pass, 0 broken links, build clean — fix regressions immediately after merging dev
• Next candidates: Serialization (index §39, partially covered — protobuf/Avro/Cap’n Proto comparison page would unify scattered content); API versioning is already covered; Quantum Computing (§35) is low-priority.

NEWLY COMPLETED (2026-08-15 — Parallel Agent Waves 1-4)

Four waves of parallel agents (7+7+7+4 = 25 agents) closed all 25 previously-missing section-level pages, covering all 50 sections of src/index.md. Each new page is a section-level integrator covering that section’s topics, with Mermaid diagrams, comparison tables, interview questions, and authoritative references.

Total new content: ~11,500 lines across 25 new pages.

All 50 sections of index.md now have dedicated coverage. Full validation suite passes: 0 broken links, SUMMARY navigation OK, MathJax balanced, Mermaid 100%.

Research expansion — round 17 follow-ups (2026-10-18)

Candidate round-18 topics (all non-overlapping with the current 129 chapters):

• Follow-up Tree: load balancers as interview ladders, database replication lag chains, Kafka consumer internals deep chain, API gateway auth chains, caching warm-up/cold-start chains.
• Assumptions Hunt: API rate-limit headers, SSL certificate rotation windows, git submodule assumptions, locale/number formatting, regex input-length assumptions, logging PII assumptions.
• Connection Atlas: feature flags and deployment coupling, database migrations and deploys, ML inference caching chains, config management pipelines.
• Why-Not Handbook: GraphQL vs REST, polling vs webhooks (mechanism focus), CI monorepo vs multi-repo, ELT vs ETL, batching vs streaming.
• Failure Casebook: only with verified primary sources (keep the contract); candidates need first-party postmortems before writing.
• Misconceptions: “silent data corruption is impossible”, “regression suite green means no bugs”, “backups mean recovery”.

Research expansion — round 20 follow-ups (2026-10-18)

Candidate round-21 topics (all non-overlapping with the current 234 chapters):

• Follow-up Tree: OAuth2 grant-flow chains, B+tree page-split chains, Kafka consumer-group rebalancing chains, browser rendering pipeline chains, payment idempotency chains.
• Assumptions Hunt: timezone/DST display assumptions, JSON number precision assumptions, disk-space exhaustion assumptions, library thread-safety assumptions, case-insensitivity assumptions.
• Connection Atlas: push-notification delivery, streaming exactly-once pipeline, service-mesh traffic management, data-lake table-format evolution.
• Why-Not Handbook: polling vs webhooks (mechanism focus, still open from the r17 list), JSON vs protobuf, replicated vs distributed SQL, structured vs unstructured logging.
• Failure Casebook: only with verified primary sources (keep the contract); candidates need first-party postmortems before writing.
• Misconceptions: “staging equals production”, “cookie vs localStorage security”, “HTTP/2 always faster”, “cold starts only affect serverless”.