Developer Builds 'Avery', an English-Like Programming Language for Beginners

·2 views

A developer has created Avery, an open-source programming language designed to make code read as naturally as plain English. The language aims to reduce syntax complexity so learners can focus on problem-solving rather than language mechanics. Avery includes common functionality such as random number generation as built-in features, eliminating the need for import statements. The project is still in early development, with some features incomplete and syntax subject to change. The creator has published Avery publicly on GitHub and is inviting feedback, particularly on syntax design and built-in functionality.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Laravel Package P2Flux Lets Buyers Pay USDC Fees Without Holding ETH

A Laravel package called P2Flux enables merchants to accept USDC payments on the Base blockchain without requiring buyers to hold ETH for network gas fees. Normally, USDC transactions on Base still require ETH to cover network costs, creating a barrier for users unfamiliar with crypto. P2Flux solves this through a payment-token sponsorship mode, where a relayer supplies the necessary ETH while the buyer pays the equivalent cost in USDC. The package is non-custodial, meaning payments settle directly to the merchant's wallet rather than being held by P2Flux. Compatible with Laravel 12 and 13, it installs via Composer and requires no API key or additional migrations, routes, or controllers.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer builds offline tool to help engineers understand AI-generated codebases

A developer has created Heimdall, a local architecture visualization tool designed to help engineers maintain understanding of codebases increasingly built by AI coding agents. The tool converts a codebase into an interactive architecture map, showing files, modules, dependencies, and component relationships without requiring users to read every line of code. Unlike many code-analysis tools, Heimdall runs entirely offline and does not use any AI models, cloud services, or external servers, keeping source code private on the user's machine. The creator argues that as AI agents generate more code faster, developers risk losing their mental model of how a system is structured — a gap Heimdall aims to fill. The project reflects a broader concern that the bottleneck in software development may be shifting from writing code to understanding what has been built.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer builds personal study system to ensure AI assists learning without replacing it

A developer has shared a structured self-study workflow designed to prevent AI tools from short-circuiting genuine understanding. The system organises learning into four sequential stages — fundamentals, academic theory, certification, and practical implementation — so concepts build on each other rather than remaining isolated. Three study modes (minimum, standard, and deep) allow the workflow to adapt to varying energy levels without breaking continuity on low-productivity days. Two AI tools are assigned distinct roles: one handles context recovery and note organisation, while the other is used interactively to challenge explanations and probe knowledge gaps. The core rule of the system is that an output only qualifies as learning if the person can explain, reproduce, and verify the reasoning behind it.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer builds CI tool to detect when Claude Code silently stops following your rules

A developer discovered that their 400-line Claude Code configuration had no automated tests, meaning model updates could silently break custom behaviors without any alerts. To address this, they built an open-source tool called config-drift-checker, which converts CLAUDE.md files, skills, and hooks into eval test cases that run on every Claude Code release or configuration change. While testing the tool, the developer intentionally sabotaged their own setup by altering a skill's trigger description, which caused the suite score to collapse from 1.00 to 0.36 and one key test to score zero across all runs. The tool uses a three-run-per-case method with history-based noise bands to distinguish genuine rule failures from random model variability, escalating to a red alert only when failures persist across consecutive runs. The tool runs as a GitHub Action and is free to use with a Claude Pro or Max subscription, with the source code available on GitHub.

0 comments Read more at DEV Community