Review dashboard (background research)
This page tracks the background research. An AI (Google Gemini) wrote four guides about Assembly and Fortran. The guides helped start the idea. The review below found their errors and gaps. Its purpose is to show areas that you did not think about. The learning itself starts on the home page.
Documents
15
AI-written guides under review
Document parts
168
21 not yet challenged
Findings
310
57 critical · 129 major · 124 minor
Awaiting source check
72
status: verify
Resolved
0
of 310 findings
0% of findings resolved (confirmed, disputed, fixed or won't fix). Update them in thefindings explorer.
Document scorecard
| # | Document | Goal met? | Confidence | Findings | Headline | Review |
|---|---|---|---|---|---|---|
| 01 | Career report Which language to learn in 90 days for career and skills? | Partly | Medium-low | 18 (3 critical) | Broken salary data; LANL claim cited to a Freelancer page; milestones not like-for-like | Pass 1 |
| 02 | Retro game report Which language to learn in 90 days to build an 80s/90s-constrained game? | No | Low | 20 (5 critical) | Its Fortran advantages don't exist on 80s/90s DOS; history contradicts its genre mapping | Pass 1 |
| 03 | Career advisor page · original ↗ Same as 01, as a quiz, roadmap and charts | No / Partly | Low | 13 (4 critical) | Quiz only restates answers and gives a verdict after one click; salary chart changes its source's numbers | Pass 1 |
| 04 | Retro advisor page · original ↗ Same as 02, as a quiz, roadmap and charts | No | Low | 10 (5 critical) | FORTRAN 77 fixed-/free-form contradiction; impossible 'locked 60 fps'; invented charts | Pass 1 |
| 05 | Assembly video report What does the video "Assembly Language in 100 Seconds" (Fireship) teach, and is this breakdown of it correct? | Partly | Medium | 23 (2 critical) | The code walk-through is correct, but the report cites the video for many claims that the video does not make. | Pass 1 |
| 06 | Assembly video page · original ↗ What does the video "Assembly Language in 100 Seconds" (Fireship) teach, and is this breakdown of it correct? | Partly | Low | 22 (3 critical) | The simulator is a useful start. But the page invents the chart and the hexdump, and the chart does not draw on load. | Pass 1 |
| 07 | Fortran video report What does the video "FORTRAN in 100 Seconds" (Fireship) teach, and is this breakdown of it correct? | Partly | Medium-low | 22 (3 critical) | The Fortran advice is mostly good, but the citations do not support the text, and the report misses most errors of the video. | Pass 1 |
| 08 | Fortran video page · original ↗ What does the video "FORTRAN in 100 Seconds" (Fireship) teach, and is this breakdown of it correct? | Partly | Low | 20 (3 critical) | The page shows invented chart numbers, and its fake compiler prints the wrong output. | Pass 1 |
| 09 | E04 benchmark report What does the video "What's the FASTEST Computer Language? C++ vs Fortran vs Cobol: E04" (Dave's Garage) teach, and is this breakdown of it correct? | No | Low | 21 (4 critical) | The report contradicts the results of the video and takes most numbers from a later forum test. | Pass 1 |
| 10 | E04 benchmark page · original ↗ What does the video "What's the FASTEST Computer Language? C++ vs Fortran vs Cobol: E04" (Dave's Garage) teach, and is this breakdown of it correct? | No | Low | 21 (4 critical) | The page invents its headline numbers and shows COBOL 50 times slower, but the video shows COBOL equal to Fortran. | Pass 1 |
| 11 | E01 benchmark report What does the video "E01: What is the FASTEST Computer Language? 45 Languages Tested!" (Dave's Garage) teach, and is this breakdown of it correct? | Partly | Low | 18 (3 critical) | The report names the right languages, but it invents results and reverses the Ada against Pascal outcome. | Pass 1 |
| 12 | E01 benchmark page · original ↗ What does the video "E01: What is the FASTEST Computer Language? 45 Languages Tested!" (Dave's Garage) teach, and is this breakdown of it correct? | No | Low | 22 (4 critical) | The page shows invented speed data as real results, and its safety simulator computes nothing. | Pass 1 |
| 13 | SNES research report What does the video "It Took Every SNES Hardware Trick To Make My Game" (Inkbox) teach, and is this breakdown of it correct? | Partly | Medium-low | 26 (3 critical) | The hardware facts hold up, but the segment map puts the audio section 28 minutes late, and the two headline corrections correct a news article, not the video. | Pass 1 |
| 14 | SNES breakdown report What does the video "It Took Every SNES Hardware Trick To Make My Game" (Inkbox) teach, and is this breakdown of it correct? | No | Low | 27 (6 critical) | The report reverses the sprite fix, the sound design and the parallax, invents a wrong address formula, and cites a news article for the video. | Pass 1 |
| 15 | SNES explorer page · original ↗ What does the video "It Took Every SNES Hardware Trick To Make My Game" (Inkbox) teach, and is this breakdown of it correct? | No | Low | 27 (5 critical) | The simulators teach designs that the developer rejected, the address calculator halves every address, and every chart number is invented. | Pass 1 |
Heatmap of the document parts
critical majorminor not yet challenged(colour = worst finding, number = finding count)
Assembly-Fortran-Comparison.md22 segments
Assembly-Versus-Fortran-Comparison.md20 segments · document-level: D02-M10
Video_Breakdown_and_Concepts.md25 segments
Status
- Divide the documents into document parts (56)
- Adversarial review, pass 1 (desk review)
- Put the findings into JSON and map them onto document parts (61 findings)
- Build the review-tracker website into
serve/ - Make the public GitHub repository and start Pages
- Replace the Actions deployment with local hooks and a
gh-pagesbranch - Decide the primary goal: learn both, for use and for fun, through a constrained game that you can play on this site (D11)
- Choose how the two languages share the game (decision O7), then the game concept and limits (O8)
- Move the site to Astro + Starlight, built and deployed by GitHub Actions (D10, D12)
- Make the language lens the center of the site (D16)
- Write all documents in ASD-STE100 (D17)
- Make the first test of the game: the two sides agree on 256 of 256 values (D19, D20)
- Review the eight video guides: 71 document parts, 169 findings, each part linked to its source video (D21)
- Review the three SNES video guides (documents 13 to 15): 41 document parts, 80 findings, 34 parts linked to a moment of the video (D25)
- Position the eight SNES learning units on the segment pages, and add the laboratory words to the dictionaries (D26)
- Test 4 of the game: the decimal counter and its wrong twin without DAA (D27)
- Tests 5 and 6 of the game: the row-start table and the terrain (PLAN.md, Phase 4d)
- Decide the PC speed that the lander targets (decision O9)
- Browser-run Assembly and Fortran demos (PLAN.md, Phase 4)
- Pass 2: get and examine the high-risk citations (the checklist is at the end of each pass-1 review):
- the LANL Fortran report,
- the dependency of OpenCoarrays on MPI,
- the release status of LFortran,
- the Microsoft FORTRAN 5.x graphics library,
- the programming language of Elite, Frontier: Elite II and M.U.L.E.
- Pass 2: challenge the 15 "not yet challenged" document parts
- Change the remaining document(s) to agree with the chosen goal
Log
| Date | Step | Output |
|---|---|---|
| 2026-09-27 | Divided 4 documents into 56 units with review worksheets | segments/ |
| 2026-09-27 | Moved documents and document parts into review/, then documents into review/segments/ (by user) | none |
| 2026-09-27 | Adversarial review, pass 1 (no sources fetched) | adversarial-review-pass1/ |
| 2026-09-27 | Wrote the process and status | README.md |
| 2026-09-27 | Put 61 findings into JSON. Mapped them onto document parts. Wrote a comparison report | findings/ |
| 2026-09-27 | Built the review-tracker static site and GitHub Pages workflow | site/, serve/, .github/ |
| 2026-09-27 | Published public repo and started GitHub Pages (Actions) | https://kairin.github.io/000-111-learn/ |
| 2026-09-27 | Removed the Actions workflow. Local git hooks built the site, and gh-pages served it | .githooks/, site/regenerate.sh |
| 2026-09-27 | Wrote the plan (with in-browser Assembly/Fortran), the decision log, and the session record | plan/ |
| 2026-09-27 | Goal decided: learn both through a constrained game that you can play in the browser. Site moved to Astro + Starlight on GitHub Actions. Hooks and gh-pages removed | site/, .github/, plan/ |
| 2026-09-27 | Language lens (A side, B side), dictionaries, STE writing standard. Review moved to background research | learn/, site/ |
| 2026-09-27 | Decisions D19 and D20. First test of the game: Fortran table, 8086 program in js-dos, check in DOSBox | game/, site/ |
| 2026-09-27 | The A-side cassette photo on all A-side pages and in the link preview. Pull requests now run the build and the game check | learn/images/, site/, .github/ |
| 2026-09-27 | The B-side cassette photo on all B-side pages. The link preview shows the two cassettes | learn/images/, site/ |
| 2026-09-27 | Split and reviewed eight video guides (documents 05 to 12). Each part shows its source video and moment | segments/, adversarial-review-pass1/, findings/ |
| 2026-09-27 | Game page: A and B panels in step with the dot; program keys for speed, pause and step | game/, site/ |
| 2026-10-05 | Added three SNES video guides (documents 13 to 15) and split them into 41 document parts | segments/ |
| 2026-10-05 | Adversarial review, pass 1, of documents 13 to 15 against the captions: 80 findings, with the moment of each part in the video | adversarial-review-pass1/, findings/ |
| 2026-10-05 | Consolidated the three reviews into eight learning units. Put each unit on a segment page. Planned tests 4 to 6 for the laboratory. Added 21 entries to the dictionaries | learn/, plan/ |
| 2026-10-06 | Test 4: the decimal counter (10,000 of 10,000 values agree) and test 4b without DAA (9,990 wrong values). Each page shows the path from the review to the test | game/, site/, plan/ |