The page now leads with adding a picture, then its answer: how wide to print it to scan from a table, for every picture whose code's version a scanner reported or the checks estimate. Below, as the evidence, all 6,499 codes sit on one scale from easy to hard.
The pixel check now takes its limits from the grid behind the print widths, in its good-light scene, by the same rule as the other checks: Fine from 4.45 camera pixels per square at every version, and a Problem limit that rises with the version.
At a slant, or on a can or bottle, the width grows so that the squares the phone sees narrowest keep the need. No width is given turned past 55°, for a code wider than the diameter over 1.2, or where the code and its margin, turned, would no longer fit the camera's view.
Behind the + under the answer, What a phone sees draws your picture as a phone's scanner would get it, printed at the width worked out and seen from its distance. It is an illustration, not a scan.
A test sheet, to pri…
2 days ago
PNG files that say how they were made, checked in your browser
A PNG file can now say how it was made, in a text chunk keyed `qr-made` that a code generator writes. Your browser reads it and checks every claim it can against the reading; one mismatch, and the whole record is ignored, with one line saying why. A file whose own details check out is marked as the original graphic by itself, and the page gives the picture's own width, explains its art in words and compares the squares meant with the squares zxing-cpp read; every number still comes from the reading. The record is removed before the picture is sent, so the server's answer is the same with or without it.
The README explains this beside the code's facts as made. Its Privacy section now says that if a file says how it was made, your browser reads that and removes it before the picture is sent; that anything else a file carries is sent with it and deleted unread, a photo's location among them; and that the page marks a file whose details check out by itself, which starts a design test that…
3 days ago
Facts from each code at its own size; the design test's server, not yet on the page
- Where zxing-cpp read a content more than once in a frame, the facts and checks leave out its reads from its own smaller copies of that frame (README, /api/measure).
- The design test's server (/api/design) is added and starts with the Space; the page doesn't offer it yet.
- Chrome on Android leaves its own scanner out of a test whose raw bytes aren't UTF-8 text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HykRoFZnnb98KRf6Xxomka
Fix styled-file hints, and add inches, copies and the many-codes rule
Styled files now get the 'looks like a file' hint, and flat files no longer get the 'looks like a photo' one. Every width and distance is also given in inches, and a printed width can be typed in inches. A picture that repeats one code gives each code's width and says how many copies zxing-cpp read. A file whose version no scanner reported is planned from the checks' estimate, and says so. The margin check reads small files better, and a margin the picture's edge may have cut is left out of the code's facts. On a picture where zxing-cpp reads 100 codes or more at one size, OpenCV's newer detector isn't run at that size or smaller, and the confidence is at most medium. OpenCV's reads of contents that aren't UTF-8 text now count. /api/measure adds copies, not_run, quiet.capped and content.sha256s.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HykRoFZnnb98KRf6Xxomka
5 days ago
Add print widths for a scanning distance, pairing, and a flag for styled files
Each read picture now leads with how wide to print its code, or a file itself, to scan from a chosen distance, worked out in the browser from the measured scanner needs and an assumed phone camera. Pictures of one code are paired in the browser by a digest of their content, and a file that already uses some of its error correction is flagged. /api/measure adds uec, uec_frame and content.sha256.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HykRoFZnnb98KRf6Xxomka
7 days ago
Fix small issues: shown-number statuses, hoverable tips, right-pile line
Each picture check now takes its status from the number the page shows, so one shown number never has two statuses. Hover tips stay while the pointer is on them. A redrawn section brings its heading into view. A picture none of the five scanners read is described by the measured codes like it. The page's remaining sentences come from copy.json. Try again sits under a failed row's text. The reading code is split into smaller modules, unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HykRoFZnnb98KRf6Xxomka
Each picture now gets a level (Easy, Everyday, Tough or Advanced) with a confidence, and checks of what the picture shows, with explanations and advice.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HykRoFZnnb98KRf6Xxomka