References

Books

Sippu, S., & Soisalon-Soininen, E. (2015). Transaction Processing: Management of the Logical Database and its Underlying Physical Structure. Springer. Chapter 6 ("Logical Database Updates") gives the cleanest treatment of the no-op-update / no-op-delete rule that governs txid allocation here.
Bernstein, P. A., Hadzilacos, V., & Goodman, N. (1987). Concurrency Control and Recovery in Database Systems. Addison-Wesley. Chapter 5 on multiversion concurrency control is the source of the "tombstone with deleted-at txid" representation we use.
Hellerstein, J. M., Stonebraker, M., & Hamilton, J. (2007). Architecture of a Database System. Foundations and Trends in Databases, 1(2). Provides the layering vocabulary (storage manager, access methods, query processor) we slice through here.

Reed, D. P. (1978). Naming and Synchronization in a Decentralized Computer System. MIT/LCS/TR-205. The original MVCC paper.
Bernstein, P. A., & Goodman, N. (1981). Concurrency Control in Distributed Database Systems. ACM Computing Surveys 13(2). Lays out the timestamp-ordering protocols that motivate our monotonic txid.

SQLite VDBE specification — https://sqlite.org/opcode.html. We do not implement a VDBE in db-15, but the opcode list is the canonical decomposition of the four operations this lab supports.
SQLite file format — https://sqlite.org/fileformat.html. The page-level layout we do not model here. Useful contrast for the wire format in CONCEPTS.md.
Standard splitmix64 reference — https://prng.di.unimi.it/splitmix64.c. All three ports use these constants verbatim.

Google's protobuf canonical encoding spec — https://protobuf.dev/programming-guides/encoding/. The discipline of sorting map entries before serialisation comes from there.
CBOR deterministic encoding (RFC 8949 §4.2). Same idea applied to a different format. Useful background for why we sort the secondary index lexicographically rather than by insertion order.