web-development

Minification, uglification, and bundling get lumped together as "build steps," but each does something distinct. Here is exactly what each one changes in your code, with a before/after size example.

The same API endpoint can return JSON, XML, or HTML depending on one request header. Here is how Accept and Content-Type headers negotiate response format, and what a 406 actually means.

Retrying a failed GET is harmless. Retrying a failed POST can double-charge a customer. Idempotency keys are the mechanism that makes retrying a payment request safe.

Auto-increment IDs leak your record count. Random UUIDs wreck index locality at scale. Here is a framework for choosing a primary key strategy based on how distributed your system actually is.

Base64 is not encryption. A hash cannot be reversed to recover a password. Mixing up hashing, encoding, and encryption is one of the most common — and most dangerous — mistakes in application security.

A 429 Too Many Requests response is not a bug — it is the API telling you exactly how to behave. Here is how rate limiting algorithms work and how to build a client that respects them.

Regex, JSON, and SQL all share one property that makes them hard to debug: dense syntax that becomes unreadable the moment it is minified, nested, or one character off. Here is why formatter tools outperform print-statement debugging.

Most real projects touch all three config formats in one pipeline — Kubernetes YAML, REST JSON, legacy XML. Here is where type fidelity breaks when converting between them, and how to validate at every stage.

Autocomplete suggests a line you read before accepting. An agent makes twenty changes and runs commands. The difference is where verification moves to — and that is now the real cost.

Cached input tokens bill at a tenth of the standard rate. Here is exactly what gets cached, the 1,024-token minimum, why writes cost 1.25x, and the prompt-ordering mistakes that silently destroy your hit rate.

Markdown fences, truncated objects, refusals, trailing prose and schema-valid nonsense — the failure modes specific to model-generated JSON, and a checklist for diagnosing each one quickly.

Input tokens, output tokens, cached tokens and cache writes — what each one costs on GPT-6 Astra, what actually counts toward your input bill, and three worked examples showing where the money really goes.