Course introduction

This is not “Theory of Database”

Same course number (CSE 532), different course
Systems course on database design and implementation
Papers from production systems (Spanner, TiDB, MongoDB, CockroachDB, Postgres, RocksDB, …)
No labs this year

Official: data structures and algorithms; a prior database course; system-level programming
You will read a systems / database paper for almost every lecture
Concurrency experience helps (transactions, replication)

Paper-heavy: exams are on the papers
Take it with caution, do not take this class if:
- you are “underload” and you do not have much time to read
- you expect an automatic C
- you have no database or systems background
Some think this is an easy course
- no labs
- lectures and exams are based on the papers
- the project is a workshop / poster / demo

Exams (60%)
- Three in-class exams: Sep 28, Oct 28, Dec 2
- We keep the two best scores (30% each)
- You may skip one exam
- No makeup tests (a miss is a zero; the drop covers it if you take the other two)
Presentations (10%)
- Required student paper presentations (not voluntary)
- Dates on the schedule; topics TBA
- Prepare the slides and send them to me a week before class
- Grading is totally subjective
Project (30%)
- Teams of about 3
- Imagine you are submitting a workshop paper or a poster to a conference
- Usually 2–6 double-column pages, plus a presentation
- Should land on one of these:
  - a useful, impactful tool
  - an interesting / cool research idea
  - a reexamination or evaluation of existing ideas or projects, with different conclusions
- Grading is totally subjective
- AI is encouraged on the project, not on exams
- Lightning talks Oct 14; poster session Dec 7
Subjective ratings (presentation and project)
- Excellent: 10 / 10
- Good: 8 / 10
- Meeting expectation: 6 / 10
- Need improvement: 4 / 10
- Not acceptable quality: 2 / 10
- Missing submission: 0 / 10
Grading standard
- A: achieve >= 90 in score, or ranking 10%
- A-: achieve >= 85 in score, or ranking 25%
- B+: score >= 80, or ranking 35%
- B: score >= 75, or ranking 50%
- B-: score >= 70, or ranking 65%
- C+: score >= 65, or ranking 80%
- C: score >= 60, or ranking 95%
- F: score < 60 and last 5%
Other bonus
- reporting a technical error I made (lecture) gives you 1 point, up to 20 points

The work that you turn in must be yours
- Project: your team’s work; do not copy another team
- Exams: work alone
You must acknowledge your influences (papers, code, people)
- Material from lecture does not require citation
Exams: no collaboration and no AI
- Do not post exam questions on the Web
AI is allowed on the project; you still have to understand what you submit
Do not publish exam solutions or another team’s project
If there are inexplicable discrepancies between exam and project contribution, we will over-weight the exam and interview you

Violate policy -> F, report to the department
- We are serious: in 19fa we caught ~20 students, and they all failed.
If you find a grading error, tell us
- Do not negotiate for extra points or extra work