92 articles
|
20 personas
|
12 categories
Free & open access — no account required
Best Practices
Library
A free, public catalog of opinionated software engineering best practices. Written to be actionable by practitioners and the AI personas that work alongside them.
How it works
Practices have enforcement levels
Every article carries a default enforcement level. AI personas configured in PushBackLog inherit these defaults — and teams can override them per persona to match their standards and risk tolerance.
Hard
Non-negotiable. The persona treats this as a blocking concern and will not proceed past a violation without surfacing it explicitly.
Soft
Strong guidance. The persona follows this unless there is a clear, documented reason not to — and must explain any deviation in its output.
Advisory
Background knowledge. The persona is aware of the practice and applies judgment — surfacing it as a suggestion when relevant.
Browse by category
12 practice areas
SOLID Principles 5
Five design principles for maintainable object-oriented code.
Clean Code 8
DRY, KISS, YAGNI, and the craft of readable, honest code.
Testing 12
TDD, BDD, the test pyramid, and confident delivery.
Security 10
OWASP, input validation, secrets, and least privilege by default.
Architecture 11
12-Factor, separation of concerns, layered systems.
Delivery 11
CI/CD, definition of done, acceptance criteria, continuous flow.
Performance 9
Lazy loading, caching, N+1 prevention, and async patterns.
Observability 6
Structured logging, distributed tracing, SLOs, and alerting.
Accessibility 3
WCAG 2.1 AA, semantic HTML, and ARIA landmarks.
Design 3
Design systems, user-centred thinking, responsive layouts.
Infrastructure 8
Infrastructure as code, container strategy, and platform hygiene.
Management 6
Engineering metrics, technical debt, and continuous improvement.
Personas
20 fictional practitioners
Each persona has a best practices profile spanning all 12 categories, with per-article enforcement levels and a rich backstory grounding their perspective.
About
What is PushBackLog?
PushBackLog is an AI-powered backlog management platform that refines work items until they can be completed autonomously — by AI personas working on your codebase in the background.
Bad tickets cost teams more than bad code. PushBackLog ends poor backlog quality at every stage of delivery — from first idea to shipped feature and back again.
This library is PushBackLog's public knowledge layer — the practices that inform how AI personas reason, plan, and execute work. It is free and openly accessible to everyone.