July 15, 2026
English class, but make it code drama
Show HN: E– – A language you dial between English and Python
This "plain English" coding idea has people asking: why not just use Python?
TLDR: E-- is a new tool that turns tightly controlled English into Python so AI only helps at setup time, not while the program runs. Commenters were torn between calling it a clever safety idea and roasting it as a fancy detour for people who could just write Python instead.
A new project called E-- wants to turn a carefully controlled form of English into Python, the popular coding language, with the big sales pitch being: let an AI help before the program runs, then lock everything down so the final result is stable and predictable. In theory, it’s the best of both worlds — human-friendly wording up front, boring reliability afterward. But in the comments, the crowd was immediately split, and honestly, that’s where the real fireworks started.
The loudest reaction was basically: "Isn’t this just extra steps?" One commenter flat-out said their first instinct was, I’d rather write Python, which is about as devastating as it gets in a coding thread. Others pushed the same vibe: if an AI writes code once and you save it, why is this system meaningfully different? That skepticism became the thread’s central drama — supporters see a neat safety rail for AI-assisted coding, while critics see a dressed-up command language pretending to be English.
Then came the identity crisis. One user pointed out the name clashes with the older and influential E programming language, while another escalated the joke with: "next: a programming language named 码". And tucked inside the snark was one genuinely curious twist: could this actually become a super-readable testing language, like a storybook for software? So yes, people mocked it, questioned it, and side-eyed the branding — but they also couldn’t stop kicking the idea around, which is usually how you know a nerdy launch has landed.
Key Points
- •E-- is a canonical-English programming language that compiles deterministically to Python.
- •The article describes a two-stage pipeline: optional LLM-based normalization or slot resolution at transpile time, followed by deterministic parsing into Python.
- •LLM use is optional, never occurs at runtime, and `{{ ... }}` value resolutions are cached to support reproducible builds.
- •E-- is distributed on PyPI as `e-minus-minus`, includes a CLI and programmatic `transpile()` API, and supports `.emm` source files.
- •The project is licensed under Apache License 2.0, and the article states that generated Python output is owned by the user and not encumbered by the project's license.