- Entry date
- 15 August 2025
- Category
- Tooling
- Access
- 🔓 Public
Latin, Tamil, and Chinese scripts, live in one UI, from day one — internationalisation done as architecture, not an afterthought.
- 4
- languages supported: English, Malay, Tamil, Chinese
- 3 scripts
- Latin, Tamil, and Chinese rendered in one UI
- i18next
- internationalisation framework, wired from the start
Honest evaluation
Deployed and in production use across all four languages/three scripts; the real proof was the multi-script layout complexity the day-one architecture absorbed, not just a string table.
What would prove or disprove it further
What would prove or disprove it further: adding a fifth language with a structurally different script (right-to-left text, for instance) without a layout rework would extend the proven claim; needing to touch core layout code to add it would show where the day-one architecture's assumptions actually stopped.
The evidence — full reasoning behind the verdict
Verdict: proven. The claim was that architecting internationalisation in from day one — for four languages across three scripts (Latin, Tamil, Chinese) in one interface — beats a late "translate the strings" pass. A working, deployed, production system rendering all three scripts correctly and consistently is real, direct evidence for that.
The proof is in what the harder problem turned out to be: not the string table, but multi-script layout — line height, font stacking, and component width for scripts with very different average character widths. Those are architectural decisions that are cheap to make once, early, and expensive to unwind later across every screen, and the report is honest that this is exactly the complexity a single-language build would never have surfaced at all.
The report also scopes itself carefully: this is a real, useful, narrow-scope system, not a research contribution — the AI components assist guidance logic already in scope rather than representing a novel technique. The interesting result is the i18n discipline, and that's the claim actually being tested here.
Internationalisation usually gets bolted on late, as a translation pass over English strings. This system took the opposite approach: four languages and three scripts, designed in from the beginning.
What it was
A multi-language parking-guidance system with building-management features and AI components, built to support English, Malay, Tamil, and Chinese in a genuinely multilingual market from day one.
What we built
Shipped & measuredA working system on React 18.3, TypeScript, Vite, and Tailwind CSS, with Zustand for state and Framer Motion for interaction polish. i18next carried the translation layer, but the harder part wasn't the string table — it was making Latin, Tamil, and Chinese scripts all render correctly, legibly, and consistently inside the same component system, across text direction, character width, and font-fallback differences that a Latin-only design never has to think about.
What we learned — including the honest negative
Doing i18n from the start, rather than retrofitting it, paid for itself directly: multi-script layout constraints (line height, font stacking, component width for scripts with very different average character widths) are architectural decisions, and they're far cheaper to make once, early, than to unwind later across every screen. This wasn't a "translate the strings" afterthought — it surfaced real UI complexity that a single-language build never would have found.
The honest edge: narrow, real, useful scope — a genuinely multilingual product built correctly, not an ambitious research bet.
Where it went / status
Working and deployed, in production use supporting all four languages. It sits alongside a much older, stewarded parking-adjacent system in the same domain window — one built stack-first-and-modern, the other maintained deliberately on an older stack — as two different honest answers to two different problems.
What is still open — kept visible
The honest edges, next to the wins. This is what turns 🔬 into 🟢 — honestly.
- A real but narrow-scope system — the AI components assisted the guidance logic and building-management features, not a novel research contribution.
Sources
- vocabotics project audit — Parking Guidance & Building Management (i18n, 4 languages), Aug 2025vocabotics internal project history · as of August 2025