How these tables are built
A reference is only worth the provenance behind it. This page says where each table comes from, which ones have been read against their primary source and which have not, and how to tell us we got something wrong.
The rule about check dates
A check date on this site means one thing: somebody opened the primary source, read the actual values, and compared them row by row. It is never set automatically to the build date. A site that stamps today's date on every page every time it deploys is not telling you when the data was checked, it is telling you when a computer ran, and readers and machines both learn to ignore it.
The consequence is uncomfortable and deliberate: most tables here have no check date. Where that is the case, the source line under the table says so in plain words instead of implying a verification that never happened.
A second rule follows from the first. Where a source could not be read literally, no number, date, or value taken from it is presented as sourced. A summary of a page is not the page. Five checked values are worth more than fifty unchecked ones, because one wrong citation discredits the other forty nine.
Verified against the primary source
4 of 15The decimal, hexadecimal, octal, and binary columns are derived from the code point by the build itself, and a test asserts that all four agree for every one of the 128 rows. The control code abbreviations are the C0 set.
Source: ANSI X3.4, mirrored in the Unicode C0 Controls and Basic Latin block.
Every registered code and its official name were taken from the machine readable registry file http-status-codes-1.csv, downloaded from IANA and read literally rather than through a summary. Unassigned ranges were dropped. Codes that are not in the registry but occur in the wild are marked as vendor or unofficial and are never presented as registered.
Source: IANA HTTP Status Code Registry.
Every named entity, and the code point it resolves to, was checked against entities.json, the machine readable file the WHATWG publishes alongside the standard. A build time test fails if any entity on the page is absent from that file or resolves to a different code point.
The uppercase and lowercase letters, their code points, and their HTML entities were checked against the WHATWG entities.json and against the code point of the character as the build reads it. The LaTeX commands are the standard control sequences; the discipline column is editorial convention, not a standard, and is labelled as such on the page.
Not independently verified
11 datasetsThese tables are the widely published sets and are believed correct, but nobody has read them back against the primary document here. In several cases the primary document is a paid standard, which is a real barrier and is named rather than glossed over. Treat these as good starting points, not as citations of last resort.
The code words and the Morse column have not been re-read against the primary ICAO and ITU documents inside this repository. They are the long standing published set, but until someone reads the source text they carry no check date.
Primary source: ICAO Annex 10, Volume II, the International Radiotelephony Spelling Alphabet.
Not re-read against the ITU recommendation inside this repository. The letter and digit patterns are the universally published set; the punctuation set is the commonly used subset and is the part most worth verifying.
Primary source: ITU-R Recommendation M.1677, International Morse code.
Shortcuts are version and platform dependent, so a sheet without a checked application version is worth less than no sheet at all. Per sheet, the version it was written against is recorded where it is known and marked as unverified where it is not. Nothing was added that could not be traced to the vendor documentation.
Primary source: The official keybinding documentation of each application.
The dot patterns have not been re-read against the UEB rulebook inside this repository. The Unicode braille cell shown next to each letter is computed from the dot numbers by the build, so the cell and the dot list can never disagree, but the dot list itself is unverified.
Primary source: Unified English Braille, International Council on English Braille.
The arm positions have not been re-read against the International Code of Signals inside this repository.
Primary source: International Code of Signals, flag semaphore.
IEC 60062 is a paid standard and was not read for this repository. The band values are the universally published set. The calculator does not add any data: it only applies the digit, multiplier, and tolerance values already in the table, so a correction to the table corrects the calculator.
Primary source: IEC 60062, marking codes for resistors and capacitors.
The ISO 4217 code list is behind a paywall and was not read for this repository. Codes and symbols are the widely published set and are unverified here.
Primary source: ISO 4217 currency codes.
The commands are the everyday subset. No flag or option was added that is not in the manual pages, and nothing destructive is listed without a warning, but the set has not been re-read against the manual pages inside this repository.
Primary source: Official Git documentation, git-scm.com.
Not re-read against the CommonMark specification inside this repository.
Primary source: CommonMark specification.
The tokens listed are the ones shared by PCRE, JavaScript, and Python. They have not been re-read against all three specifications inside this repository, which matters here because the engines differ in exactly the corners a cheat sheet tends to gloss over.
Primary source: ECMAScript language specification, regular expressions.
Roman numerals have no single governing standard for the subtractive forms in modern use. The seven symbols and the six subtractive pairs are the conventional set.
Primary source: ISO 80000-2 and common typographic practice.
How a table gets built
- The primary standard is identified first, before any data is typed. A secondary source is only used to cross check, never as the origin.
- Where the standard publishes a machine readable file, that file is used directly. The HTTP status codes come from the IANA registry CSV and the HTML entities from the WHATWG entities.json, both read as files rather than as prose about them.
- Derived columns are computed, not typed. The hex, octal, and binary columns of the ASCII table are calculated from the code point, and the Braille cell is calculated from the dot numbers, so a derived column can never disagree with the value it derives from.
- Editorial columns are labelled as editorial. Where a table explains what a code means in plain English, or which codes get confused with each other, that is written here and is not attributed to the standard.
- Nothing that destroys data appears without a warning next to it, and no example is written to be pasted blind.
How often it is rechecked
Standards move at different speeds, so the cadence follows the source rather than the calendar. The IANA status code registry and the WHATWG entity list change occasionally and are worth rereading yearly. ASCII has not changed in decades and needs no cadence at all. Application keyboard shortcuts change with every major release and are the fastest rotting content on the site, which is why each sheet states the version question openly rather than implying currency.
If something here is wrong
Please say so. Send the page, the row, and the source that contradicts it to corrections@quickref.tips. A correction with a citation gets fixed and recorded in the changelog, with the date. A correction without one gets checked first, which takes longer.
None of the data here is copied from another cheat sheet site. Where a value could not be traced to a standard, it is marked as unverified rather than quietly kept.