Tech Blog

Tech Blog

We share insights gained on the software development front line, including web development, algorithms, and performance optimization.

QR Code Specifications Explained (with sample QR codes)

From versions (1–40) and module counts, data capacity, and the four encoding modes to error correction, structure, and masking. We explain the specification while looking at actual generated sample QR codes. Includes a link to the generator tool.

Read more →

Barcode Specifications Explained (with sample barcodes)

We explain the characteristics and structure of EAN-13/JAN, EAN-8, UPC, CODE39, CODE128, ITF, and Codabar, along with how check digits work, using actual generated samples. Includes a link to the generator tool.

Read more →

How Top Engineers Use Claude Code

We look at how leading engineers around the world—Armin Ronacher, Simon Willison, Mitchell Hashimoto, and Boris Cherny—actually use Claude Code, drawing on their original posts.

Read more →

How Flask's Creator Armin Ronacher Uses Claude Code

claude-yolo with Docker, keeping logs in files, designing agent-friendly code, and CLI over MCP. We dig into the Flask creator's practices with quotes from the original article.

Read more →

How Django Co-creator Simon Willison Uses Claude Code

The idea of "designing the agent loop," YOLO mode and sandboxes, throwaway $5 Fly.io orgs, and using tests as a means of verification—explained with quotes.

Read more →

How Claude Code's Creator Boris Cherny Uses It

Running multiple Claudes in parallel, recording mistakes in CLAUDE.md, /commit-push-pr, moving from plan mode to auto mode, and "2–3x quality through verification"—with sources.

Read more →

How Ghostty's Creator Mitchell Hashimoto Uses AI Agents

Delegating research, harness engineering (AGENTS.md), real examples and costs in Ghostty, and the cautionary idea of "agent psychosis"—explained with quotes from his own blog.

Read more →

Running Projects with Claude Code

The Explore → Plan → Code → Commit loop, spec.md/plan.md, CLAUDE.md, git worktree, and TDD. We organize a repeatable workflow based on official docs and practitioners' blogs.

Read more →

Building Test Scenarios with Claude Code

The verification loop, RED-GREEN-REFACTOR in TDD, E2E verification via browser automation, and cross-verification with sub-agents. We explain how to build a test strategy based on research from external blogs.

Read more →

Team Collaboration with Claude Code

Sharing CLAUDE.md, committing the .claude/ directory, GitHub Actions integration, MCP servers (Slack/Linear), and onboarding. We introduce how to build the foundations for team collaboration.

Read more →

Documentation Management with Claude Code

CLAUDE.md and automatic memory, initialization with /init, generating JSDoc/OpenAPI from code, and keeping docs in sync with hooks and CI. We explain how to keep documentation from going stale.

Read more →

Understanding QR Code Error Correction Levels — Why They Still Scan When Dirty

We explain, in clear terms, how QR codes remain readable even when partly hidden by a logo or smudged, from the perspective of error correction levels (L/M/Q/H) and Reed-Solomon codes.

Read more →

Intro to Sorting Algorithms — 7 Methods Chosen by Complexity

From bubble sort to quicksort and merge sort. We organize how the major sorting algorithms work and their time complexity (Big O notation), and explain how to choose the right one for the job.

Read more →

Core Web Vitals: A Practical Guide — Improving Speed and Perceived Performance

Centered on the three metrics LCP, INP, and CLS, we summarize concrete measures to improve website load speed and user experience—from measurement to implementation, with a practical focus.

Read more →