The 1947 Moth That Gave Software Bugs Their Name Still Has Lessons for Engineers

·1 views

On September 9, 1947, engineers debugging Harvard's Mark II electromechanical computer discovered a moth lodged inside Relay #70, shorting its contacts and causing errors. The team, part of a U.S. Navy computing effort that included Grace Hopper, taped the insect into their logbook at 15:45 with the note 'First actual case of bug being found.' While Hopper is often credited with coining the term 'bug,' the word had already been used by engineers for decades, with Thomas Edison employing it as early as 1878. The logbook entry's humor lies in the fact that 'bug' was already abstract slang for a fault — making this one rare instance where the bug was literally real. The incident remains a practical reminder for modern embedded and IoT engineers that hardware faults, not just software errors, are often the true source of system failures.

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)

How to Migrate WordPress to Next.js Without Tanking Your SEO Rankings

Migrating a WordPress site to Next.js involves far more than fetching posts via the REST API — preserving SEO requires maintaining existing URLs, metadata, redirects, sitemaps, and images. The core principle is that nothing Google already indexes should change unless deliberately improved, since altering URLs, content structure, and templates simultaneously destroys accumulated ranking signals. A complete pre-migration inventory of every published URL, rendered content, metadata, and internal links is essential before writing any Next.js code. Standard WordPress WXR exports are often inadequate for page-builder-heavy sites, as they omit rendered HTML and key SEO fields. Tools that capture fully rendered content and structured SEO data can serve as a reliable source of truth throughout the migration process.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Solon AI v4.0.3 Introduces 'Talent' Layer to Improve Agent Tool Management

Solon AI version 4.0.3 introduces a concept called 'Talent' to address limitations of using bare function-based tools in AI agent frameworks. While standard tools only tell a model what to call and what arguments to pass, they lack context awareness, domain grouping, and procedural guardrails. A Talent bundles tools together with a standard operating procedure and activation rules, allowing capabilities to appear or hide based on the current request context. This design reduces token usage by keeping inactive domain tools out of the model's context window and improves compliance by preventing the model from acting out of sequence. Plain tools remain suitable for simple, low-risk, self-contained functions, while Talents are recommended when multi-step workflows, role-based access, or reusable domain logic are required.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Browser clock clamping exposed a gem ID collision bug hidden from JVM for months

A developer building a match-3 game discovered that the same Java source code produced 301 duplicate gem IDs in the browser but zero on the JVM across 128,000 gems. The root cause was reliance on System.nanoTime() for ID generation, which works reliably on the JVM but fails in browsers because Spectre mitigations clamp the clock to roughly 100-microsecond resolution. Running identical logic on two runtimes — the JVM and TeaVM-compiled JavaScript — acted as a free cross-check that surfaced the discrepancy without requiring the developer to anticipate the failure. The fix replaced the timestamp-and-random scheme with a simple atomic counter, eliminating any dependency on platform clock resolution. Three regression tests were added to CI to ensure gem IDs can never silently revert to clock-dependent generation.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer fixes recurring job timeouts by splitting tasks with sub-step idempotency

A developer running an automated vault-auto-ingest job faced repeated timeouts over three consecutive days when a single processing step failed to complete within its 40-minute window. The step responsible for digesting 28 hours of Claude and Codex conversation logs was too large to finish reliably during busy periods, causing a completion marker to never be set and all four daily retry slots to fail in sequence. The fix involved splitting the problematic step into two independent sub-tasks, each with its own marker file and a 25-minute timeout, so that a timeout in one would not force the other to re-run. A shared composite marker is only written once both sub-steps succeed, preserving the existing outer retry logic. This approach, termed sub-step idempotency, ensures that already-completed units of work are skipped on subsequent retry attempts within the same day.

0 comments Read more at DEV Community