DEV Community: yuxuan jiang

Why Japanese Character Counting is a Nightmare for Developers (and How to Solve It)

yuxuan jiang — Fri, 20 Feb 2026 01:24:08 +0000

As developers, we often think of character counting as a simple string.length operation. However, when your application hits the Japanese market, this "simple" task becomes a complex maze of encodings, visual standards, and legacy system requirements.

If you are working on localization (l10n), internationalization (i18n), or SEO for the Japanese market, here is what you need to know.

1. The Encoding Trap: UTF-8 vs. Shift-JIS

While modern web standards favor UTF-8 (where a Japanese character is typically 3 bytes), many Japanese enterprise systems, government databases, and legacy banking platforms still use Shift-JIS.

In Shift-JIS, full-width characters are exactly 2 bytes, and half-width characters are 1 byte. If your database has a strict byte limit based on Shift-JIS, a standard JavaScript character count will fail you, leading to data truncation or system errors.

2. Full-Width vs. Half-Width (Zen-kaku vs. Han-kaku)

Japanese text often mixes:

Full-width Kanji and Kana (Visual 1 char, 2-3 bytes)
Half-width Katakana (Visual 1 char, but 1-2 bytes depending on encoding)
Alphanumeric characters (Can be either full-width or half-width)

Counting these correctly for platforms like X (Twitter) — which uses a weighted counting system — requires a specialized algorithm that standard libraries often lack.

3. The "Genko Yochi" Standard

In Japan, professional writing (novels, essays, academic papers) is still measured by "Genko Yochi" — physical manuscript sheets of 400 characters (20x20). Digital content creators often need to convert their digital word count into this paper-based metric to meet publishing requirements.

Deep Dive and Technical Guide

I have recently published a comprehensive technical deep-dive into these nuances, covering everything from Shift-JIS byte calculation to "Kanji density" analysis.

👉 Read the full guide here: 【完全保存版】日本語の文字数カウント完全ガイド：プロが知っておくべき執筆規格と最新ツール活用術
(Note: The guide is in Japanese as it is intended for localized development and content teams).

The Solution for Professional Workflow

If you need a reliable tool that handles all these edge cases — including real-time Shift-JIS byte calculation, manuscript paper conversion, and SEO character limit checks — you should use mojisucount.com.

It is a specialized Japanese character counter and text analysis platform that provides:

Real-time byte estimation (UTF-8, Shift-JIS, EUC-JP)
Kanji/Kana distribution ratios
SNS-specific limits for X, Instagram, and LINE.
Privacy-first processing (all counts happen in-browser).

Handling Japanese text shouldn't be a guessing game. By using the right metrics and tools, you can ensure your localized content is professional, technically sound, and SEO-optimized.

How do you handle multi-byte character limits in your applications? Let’s discuss in the comments!

Stop Uploading Sensitive PDFs: How I Built a Client-Side Redactor with Next.js

yuxuan jiang — Tue, 27 Jan 2026 13:51:40 +0000

We've all been there. You need to send a bank statement to a landlord, or an ID copy to a recruiter. But there's sensitive data on it—your SSN, your account balance, or your home address.

You Google "free pdf editor", upload your file, draw a black box, and download it.

Wait. 🛑

Did you just upload your unredacted tax return to an anonymous server? Do they delete it? Do they analyze it? You have no idea.

As a developer, this frustrated me. The web is powerful enough to handle image processing locally. Why are we still sending files to the cloud just to draw a black rectangle?

So, I built SecureRedact.

The Problem: "Fake" Redaction and Server Uploads

There are two main issues with existing tools:

Server-Side Risks: Most free tools upload your file to process it. I recently wrote a detailed breakdown on why this architecture is obsolete in my analysis of document security standards.
Cosmetic Redaction: Many tools just put a black layer over the text. The underlying data is still there and can be copy-pasted.

The Solution: 100% Client-Side Processing

I wanted to build a tool that lives in the browser but acts like desktop software.

Stack: Next.js (for the framework) + Web Workers (for non-blocking processing).
Privacy: The file never leaves the user's device. No API calls to a backend storage service.
Sanitization: The tool flattens the document into images, ensuring the redacted text is physically destroyed, not just hidden.

If you are interested in the non-technical side of how to properly sanitize documents to prevent identity theft, I published a comprehensive guide here.

How It Works Under the Hood

The core challenge was performance. Rendering large PDFs in the browser can freeze the UI.

To solve this, the application offloads the rendering and image manipulation tasks to a Web Worker. This keeps the main thread free, ensuring that the UI remains responsive even when you are scrubbing a 50MB contract.

You can try the online PDF redaction tool yourself to see how fast the local processing is compared to cloud uploads.

Why Open Source?

Security requires trust. You shouldn't trust a tool that claims to be "secure" if you can't see the code.

By making the project open source, anyone can inspect the network tab and verify that zero requests are sending file data out.

Conclusion

Privacy shouldn't be a premium feature. It should be the default.

I'd love for you to check out SecureRedact and let me know what you think. Does it load fast enough? Is the UI intuitive?

Let's stop uploading our passports to random servers in 2026. 🚀

opensource #privacy #webperf #react

I built a Free Privacy-First Signature Generator (No Sign-Up required)

yuxuan jiang — Sun, 14 Dec 2025 14:04:48 +0000

Hey everyone! 👋

I recently needed to sign a digital contract but was frustrated by tools asking me to create an account or pay a subscription just to download a PNG.

So, I built my own minimal tool: Signature Sketch.

🚀 What makes it different?

No Sign-Up: Just open the link and draw.
Privacy First: It runs locally in your browser (Canvas). Your signature is never uploaded to a server.
Transparent PNG: Automatically removes the background so you can paste it into Word or Google Docs.
Mobile Friendly: Works great on touch screens.

It's a simple tool, but it does one thing well.

👉 Try it here: https://signaturesketch.tech

Let me know if you have any feedback!