FIELD NOTES / OZLIN INFO
Blog
Practical notes on AI automation, secure web delivery, cybersecurity, document processing and game-development R&D. Use the curated paths below to choose a starting point, then browse the full chronological archive.
Start with a practical path
AI automation: AI Chatbots for Australian SMEs
Document processing: Document Scanner Prototype and Reading an Older PyTorch HMER Repository Responsibly
Web and WordPress: WordPress Performance Checks
Cybersecurity: Practical Priorities for Australian SMEs
About the practice: About Ozlin Info
Public game-development R&D archive: Explore public R&D projects
-

Optimise Unity Rendering: Batching, Instancing and Profiler Evidence
Diagnose Unity rendering cost before choosing SRP Batcher, GPU Resident Drawer, GPU instancing or static batching, then verify CPU, GPU, memory and visual results on target hardware.
-

Particle Systems for Games: Build Effects to a Measured Budget
Design particle effects from a communication goal and a measured CPU, GPU, memory and overdraw budget, with scalable emitters, pooling, culling and target-device tests.
-

Cross-Platform Game Input: Actions, Remapping and Device Changes
Build input around gameplay actions, control schemes, rebinding, dead zones, device hot-plug, multiple users, glyphs, accessibility and capability-tested haptics.
-

Game State Management: Choose Ownership Before a State Machine
Separate global flow, entity, interface, save and network state; then use finite, hierarchical or stack-based state models with explicit events, guards, ownership and tests.
-

Collision Detection in Games: Broad Phase, Narrow Phase and CCD
Choose collision shapes, filtering, broad-phase structures, narrow-phase tests and continuous collision detection from game requirements and measured scene distributions.
-

Build a Stable Pygame Loop with Fixed Updates and Interpolation
Build a complete pygame-ce loop that pumps events, clamps frame time, advances fixed simulation steps, interpolates rendering, handles focus and exits cleanly.
-

C++ Game Memory: Ownership, Locality and Evidence Before Pools
Improve C++ game memory by making ownership explicit, preferring values and contiguous storage, measuring allocations and cache behaviour, and adding arenas or pools only for proven workloads.
-

Choose a 2D Physics Engine with a Reproducible Project Benchmark
Compare Box2D, Rapier 2D and an engine-integrated physics option through dated licences, feature requirements, deterministic replay, tooling and a project-specific benchmark.
-

Local SEO for Australian Service Businesses: A 2026 Reality Check
Build local search visibility from genuine eligibility, accurate business information, useful service pages, technical quality and measured Search Console evidence—not ranking folklore.
