Hasan's Journal

Stories, lessons, and scars from production.

Mehedi Hasan

Writing

Projects, case studies & notes.

What I ship, what I learn, and what I'd do differently next time.

ForMesh: Multi-Section Forms for React, Without the Re-render Tax

Most form libraries are built around one assumption: a form is a single flat object. That assumption holds for a login form or a signup form. It stops holding the moment you build enterprise software - an employee record that is Personal Info + Job Info + Bank Info + Documents, each section owned by a different part of the UI, all of it merging into one object at submit time. I ran into that gap repeatedly while building form-heavy modules for an 8-module ERP platform, and eventually stopped working around it and built ForMesh (published as react-formesh) instead: a small, framework-agnostic form-state core for exactly that shape of form. This post walks through the problem, the architecture decisions that came out of it, and where the project stands today. It's currently in active development and not yet published to the public npm registry - the API below reflects what's implemented and tested now.

#DNS#WebDev#DevOps#Architecture

Inside a Large ERP Frontend: How I Kept 8 Modules and 120,000+ Lines of Code Coherent

A case study from ONE ERP, a cloud ERP platform for garment and apparel manufacturing A large ERP rarely fails because of one big bug. It fails slowly. A form re-renders on every keystroke. A permission rule gets copied into six screens. A costing formula drifts between two modules. A release stalls because one team isn't ready. I led frontend engineering on ONE ERP at BerryLabs, an eight-module platform covering styles, costing, budgets, HR and payroll, inventory, invoicing, commercial and shipping documents, and quality inspection. This is the playbook I ended up with, and the reasoning behind each piece.

#DNS#WebDev#DevOps#Architecture

Why I Built a Form Library Instead of Reaching for One

I'm writing this partly as a build log and partly as an argument. The build log part is straightforward: I made a form state library called ForMesh, and I want to explain what it does and why it's put together the way it is. The argument part is the more interesting bit to me — I want to make the case that "just use Formik" or "just use React Hook Form" isn't always the right answer, and that sometimes the more disciplined move is to sit with a recurring problem long enough to actually understand its shape before reaching for a dependency. If you build React apps for a living, especially anything in the ERP, CRM, or internal-tools space, I think this will feel familiar. If you're earlier in your career and wondering when it's actually worth building something instead of installing it, I hope this is a useful, honest data point either way.

#DNS#WebDev#DevOps#Architecture