Algorithm Review: The Engineering Checkbox

Introduction: The “Hidden” Bug

Most bugs cause errors. Algorithmic bugs cause meltdowns. A logic that works fine with 10 test items can take 10 minutes when it hits 10,000 real production items. As a developer or reviewer, you need to develop an “eye” for these patterns during Code Review.

The Reviewer’s Checkbox

1. The Nested Loop Trap ($O(N^2)$)

• The Pattern: A for loop inside another for loop over the same list.
• Reviewer says: “Can we replace the inner loop with a Hash Map lookup to turn this into $O(N)$?”
• Example: Checking for duplicates or merging two lists.

2. The String Concatenation Trap

• The Pattern: Building a long string inside a loop using str += "...".
• The Risk: In many languages (like Java or Python), strings are immutable. This creates a new string every time, leading to $O(N^2)$ memory copying.
• Reviewer says: “Use a StringBuilder or join an array at the end.”

3. The Unbounded Recursion

• The Pattern: Deep recursion without a clear depth limit.
• The Risk: Stack Overflow errors.
• Reviewer says: “Is there a limit to the input depth? If not, can we rewrite this using an Iterative approach with a stack?“

4. The Database N+1 Query

• The Pattern: Querying the database inside a loop.
• Reviewer says: “Batch this! Fetch all IDs first, then use one query with WHERE id IN (...).”

5. The Sorting Necessity

• The Pattern: Sorting a whole list just to get the first 3 items.
• Reviewer says: “Use a Min-Heap or Quickselect to get Top K in $O(N)$.”

The Developer’s Self-Correction

Before you submit a PR, ask yourself:

• “What happens if N is 100,000?” (Scalability)
• “Is this deterministic?” (Reliability)
• “Did I use the right data structure?” (Efficiency)

Typical Business Scenarios

• ✅ CSV Export: If exporting 1M rows, don’t build the whole array in memory. Use Streaming.
• ✅ Permissions Check: Don’t check every user against every role in a nested loop. Use a Bitmask or Set.

Summary

"Code reviews aren't just about syntax; they are about protecting the system's future. An algorithmic checkbox ensures that the code you write today doesn't become the outage you fix tomorrow."