WEEK 09 · 2026.09.21-09.27 · WRITER GATE
Day 3 - News Aggregator TRANSFER: blind attempt and HLD delta
2026-09-23 · WED · 先从 FB Post Search 迁移盲做,再用 Google News 原文修正 crawler、dedupe、canonical article 和 feed materialization。
今日固定时间表
三个学习窗口必须保持 exactly 08:50-10:55, 14:30-16:15, 20:30-21:15。
| 08:50-09:25 | 08:50-10:55 system design Blind transfer attempt: reuse index/rank vocabulary, but list assumptions before reading. |
|---|---|
| 09:25-10:05 | 08:50-10:55 system design Read News Aggregator headings through HLD and three functional paths. |
| 10:05-10:35 | 08:50-10:55 system design Write exact delta: changed requirements, invalid assumptions, new bottleneck, follow-up. |
| 10:35-10:55 | 08:50-10:55 system design DDIA Ch13: derived article/feed state, materialized cache and stale derived state. |
| 14:30-16:15 | NeetCode Graphs Three contiguous Graphs slots below; no other tag. |
| 20:30-21:15 | Transfer script audio pass: explain crawl lag, dedupe lag and feed materialization lag separately. |
算法模块
NeetCode Graphs, exactly three slots today, contiguous slots 7-9 of 21.
| # | Problem / mode | Pattern | Invariant | Bug risk | Time | Space |
|---|---|---|---|---|---|---|
| 7 | Course Schedule Graphs · NEW · 30m solve + 5m evidence | topological cycle detection | a course is finishable after all prerequisites leave visiting state | not distinguishing visiting vs visited | O(V+E) | O(V+E) |
| 8 | Course Schedule II Graphs · NEW · 30m solve + 5m evidence | topological order | output order appends only nodes whose prerequisites are satisfied | reversing order in DFS incorrectly | O(V+E) | O(V+E) |
| 9 | Clone Graph Review Graphs · NEW · 30m solve + 5m evidence | memoized traversal under time pressure | the memo map breaks cycles and preserves identity | using value as key when values may collide | O(V+E) | O(V) |
今日产出
| 必须交付 | News transfer blind attempt plus complete link map first half. |
|---|---|
| 修复规则 | If the answer treats news as user posts, add crawler/fetcher/parser/deduper/canonical cluster. |
Hello Interview 精确目录
使用 live heading 的 original title 和 direct anchor。
| 项目 | Level | 精确 canonical section | Disposition | Day |
|---|---|---|---|---|
| News Aggregator / Google News | h2 | Understanding the Problem | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | Functional Requirements | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | Non-Functional Requirements | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h2 | The Set Up | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | Planning the Approach | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | Defining the Core Entities | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | API or System Interface | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h2 | High-Level Design | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | 1) Users should be able to view an aggregated feed of news articles from thousands of source publishers all over the world | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | 2) Users should be able to scroll through the feed "infinitely" | blind attempt then reading | Day 3 |
| News Aggregator / Google News | h3 | 3) Users should be able to click on articles and be redirected to the publisher's website to read the full content | blind attempt then reading | Day 3 |
DDIA 精确决策卡
Only DDIA Ch4 and Ch13 decision links are allowed.
| Day | Project | Decision | Link |
|---|---|---|---|
| Day 3 | News Aggregator | feed rows and index documents are derived from crawled source articles | https://ddia.vonng.com/ch13/#id446 |
| Day 3 | News Aggregator | homepage/category feed caches are materialized views with staleness budgets | https://ddia.vonng.com/ch13/#id451 |
Artifacts
PDF is intentionally absent; Week 9 requires MP3 only for the FULL/TRANSFER scripts.
Detailed lecture notes, audio, recall scripts, PDFs, Staff Q&A, and mock packs are archived locally and are intentionally not published on this site.