Skip to content

fix(ch07): improve MC distractors to eliminate easy-guess answers - #34

Merged
calmcoconut merged 1 commit into
mainfrom
fix/ch07-mc-distractors
Jun 30, 2026
Merged

calmcoconut merged 1 commit into
mainfrom
fix/ch07-mc-distractors

Conversation

@calmcoconut

Copy link
Copy Markdown
Owner

Summary

Audit and improvement of Chapter 7 (Sharding) multiple choice question distractors. Several wrong answer options were too obviously absurd or off-topic, making the correct answer easy to identify without domain knowledge. This PR replaces them with plausible alternatives that require genuine understanding to eliminate.

Questions Changed

Question Issue Fix
Q2 — Single-machine per-core sharding DBs Options C/D named completely unrelated DB categories (cluster DBs, cloud warehouses) Replaced with high-performance analytical DBs (Elasticsearch/ClickHouse/ScyllaDB, TimescaleDB/InfluxDB/QuestDB) that a learner might plausibly confuse
Q9 — Why language hash functions are unsuitable Option C claimed hash functions "only work on integers" — obviously false to anyone who has used hashCode() on a String Replaced with a plausible performance-based wrong answer about computational slowness
Q12 — Fixed shard count trade-offs Options C/D described impossible system behaviors (SSD/HDD forced by shard count; node-count ceilings of 2 or 10) Replaced with realistic operational concerns: coordinator failover overhead and difficulty achieving fine-grained load balancing
Q17 — Risk of automated rebalancing Options C/D described nonsensical behaviors (permanent boundary locks; DNS blocking) Replaced with real failure modes: shard-split storms and routing tier lag during rebalancing
Q18 — Routing tier strategy Option D described client-side RAM caching, which is not a routing strategy at all Replaced with a push-notification shard-aware driver description that blends options A and C
Q23 — DBs that use local secondary indexes Option C listed BigQuery/Snowflake/Delta Lake (cloud warehouses with no row-level sharding) Replaced with CockroachDB/YugabyteDB/TiDB (distributed SQL DBs that actually use global indexes, making this a genuine knowledge test); updated option A accordingly
Q28 — DynamoDB GSI write propagation Option D claimed weekly maintenance window updates — an obvious joke Replaced with a 2PC-based distractor that requires knowing DynamoDB specifically avoided this approach for performance reasons

What Was Not Changed

All correct answers, explanations, write-response questions, flashcards, and MISCONCEPTION_EXPLANATIONS are unchanged. Only the wrong-answer distractor text was modified.

…guess

- Q2: Replace Cassandra/MongoDB/HBase and BigQuery/Snowflake/DeltaLake
  with Elasticsearch/ClickHouse/ScyllaDB and TimescaleDB/InfluxDB/QuestDB
  as more believable high-performance DB distractors
- Q9: Replace nonsensical "integers-only" distractor with a plausible
  performance-based wrong answer about computational slowness
- Q12: Replace absurd SSD/HDD and node-count-limits distractors with
  realistic operational concerns about coordinator overhead and
  fine-grained load balancing
- Q17: Replace impossible "locks boundaries forever" and "blocks DNS"
  distractors with real failure modes around split-storm I/O and
  routing tier lag
- Q18: Replace "clients cache all records in RAM" (not a routing strategy)
  with a realistic push-notification shard-aware driver description
- Q23: Replace cloud data warehouse option (BigQuery/Snowflake/DeltaLake)
  with distributed SQL DBs (CockroachDB/YugabyteDB/TiDB) that actually
  use global indexes, making the distractor a genuine knowledge test;
  update option A accordingly
- Q28: Replace "weekly maintenance window" joke distractor with a
  plausible 2PC-based wrong answer
@calmcoconut
calmcoconut merged commit e050c36 into main Jun 30, 2026
1 check passed
@calmcoconut
calmcoconut deleted the fix/ch07-mc-distractors branch June 30, 2026 22:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant