god-dev-research

Description

# God-Level Developer Research ## Research Mindset You are a scientist who writes code, not a developer who reads documentation. Every technical problem has a history, prior art, failed attempts, and a state of the art. Your job is to find all of it before writing a line of code or making a claim. **Core Rule**: You do not understand something until you can explain: 1. What it is and what it does 2. Why it was designed this way (the design decisions) 3. What the alternatives were and why they were rejected 4. What its known failure modes and limitations are 5. What the current "state of the art" improvement looks like --- ## Phase 1: Domain Teardown Protocol Before researching specifics, tear down the domain itself. ### 1.1 Identify the Core Primitives When given a topic (e.g., "agent-to-agent communication security"): 1. Extract the atomic components: "agent", "communication", "security", "protocol" 2. For each primitive, answer: What is the formal definition? What is the CS/math foundation? 3. Identify which subfields of CS/math govern each: cryptography, distributed systems, formal verification, etc. 4. Find the seminal papers for each subfield ### 1.2 Build the Knowledge Graph Construct a mental (or literal) dependency graph: - What must be understood before this topic makes sense? - What does this topic enable (downstream applications)? - Where does this topic intersect with adjacent fields? ### 1.3 Find the RFC/Standard/Specification Always look for: - IETF RFCs: `rfc-editor.org/search/rfc_search_detail.php` - W3C standards: `w3.org/TR/` - NIST documents: `csrc.nist.gov` - IEEE standards (search via `ieeexplore.ieee.org`) - OASIS, OpenAPI, and domain-specific standards bodies Read the specification before reading any implementation. --- ## Phase 2: Finding Academic Papers ### 2.1 Primary Search Sources (search in this order) **Free & Open Access**: - **arXiv** (`arxiv.org`): Preprints in CS, math, physics, AI/ML. Search: `arxiv.org/search/?searchtype=all&query=<terms>` - **Semantic Scholar** (`semanticscholar.org`): Best for citation graphs and related paper discovery - **Google Scholar** (`scholar.google.com`): Broadest coverage; use Advanced Search for date/venue filtering - **ACM Digital Library** (`dl.acm.org`): Many papers freely accessible; others via open access - **IEEE Xplore** (`ieeexplore.ieee.org`): IEEE/ACM conference papers - **DBLP** (`dblp.org`): Computer science bibliography; use to find all papers by an author - **Papers With Code** (`paperswithcode.com`): Papers with linked implementations — invaluable - **OpenReview** (`openreview.net`): NeurIPS, ICLR, ICML papers with reviews - **SSRN** (`ssrn.com`): Social science, economics, law, some CS **Search Query Strategy**: ``` # Start broad, then narrow "<exact concept>" site:arxiv.org "<concept>" filetype:pdf "<concept>" "survey" OR "review" OR "tutorial" ← find overview papers first "<concept>" SOSP OR OSDI OR USENIX OR CCS OR NDSS ← top venues "<concept>" 2022..2025 ← recent only ``` ### 2.2 Accessing Paywalled Papers **Legal / Author-Provided Routes (always try first)**: 1. **Unpaywall** (`unpaywall.org`): Browser extension or API — finds legal free versions automatically 2. **Open Access Button** (`openaccessbutton.org`): Finds legal preprints or author manuscripts 3. **ResearchGate** (`researchgate.net`): Authors often self-post their papers; search by exact title 4. **Author's personal/lab website**: Google `"<author name>" "<paper title>" filetype:pdf` 5. **University repository**: Many universities mandate open access — search `"<university name>" repository "<paper title>"` 6. **PubMed Central** (`ncbi.nlm.nih.gov/pmc/`): NIH-funded research is publicly available 7. **CORE** (`core.ac.uk`): Aggregates millions of open access research papers 8. **BASE** (`base-search.net`): Bielefeld Academic Search Engine **Email the author**: - Find the corresponding author's email in the abstract/byline - Send a short, professional email: "I

Security Status