Page generated from: SET-File_OnePage_v4.5.md

---

SET-File_OnePage_v4.5.md

Set|Files – Outline Summary

SET (.set) format v4.5

Set|Files are a Human-and-AI readable data format – using strings as the fundamental data-type. Clean, clear, easy to read, easy to edit – smaller and simpler than other common formats. Set|File is the norm; Q-Set (.qset) is the same format with every default locked – see the Q-Set One-Page v4.5. A Set|File reader handles everything on this page as a matter of course; nothing is declared – if a file uses it, it is there. X-Set (.xset) adds loaded modules that bend core rules; it will not open in a plain reader – see the Specification, Appendix A.

Head – every file opens with three fixed lines

• Line 1: the filename – the content's canonical name, which need not match the name on disk: myconfig.set
• Line 2: the Conformance line – required: Conformance|v4.5 – always written with the default |
• Line 3: empty, or the delimiter map (below) – nothing else may stand there

Delimiter map (line 3, bare) – a variable table for the parser; any slot may change, brackets included

• Default: ` :[]:{}:|:\:!: ` – the first character is the separator used to read the line itself
• Slots: 1 map separator (also the per-line override prefix) · 2 Group brackets · 3 Text-Group braces · 4 primary delimiter · 5 escape · 6 secondary delimiter · 7 quote character
• Every slot holds a character – to keep a default, write the default; no two slots alike; applies from line 4 on
• Example, semicolon primary and comma secondary: :[]:{}:;:\:,: `

Parsing – NewLine (LF or CRLF), one linear pass; delimiters are known before anything else

• Trailing whitespace is removed from every line before classification; a whitespace-only line is an empty line
• Reserved first characters: [ (Group forms), : (per-line override), whitespace (in-group comment) – nothing else
• Fields are trimmed by default; Trim|off in [THIS-FILE] hands fields back exactly as written (structure unchanged)
• The file is sacred: load → read → save is byte-identical unless an edit changed something
• A reader that meets a malformed form refuses – it never guesses

Comments – any line outside a Group; no comment marker

• Header comment: comment lines touching the header. Footer comment: lines directly after [EOG]. Both optional conventions
• In-group comment: a line beginning with whitespace (with at least one visible character after it)
• In SetTag hosts only, > directly after the comment marker (//>, #>) binds the comment to the following Group across empty lines

Group(s) – [NAME] alone on its line

• Names: letters, digits, -, _; case-sensitive; unique within the file; ALL-CAPS recommended, not required
• Reserved in any letter case: THIS-FILE, EOG, EOF
• Close: empty line, next header, [EOG] at the start of a line, [EOF] / end of file
• "Table" and "key|value" describe a Group's format, not a kind of Group – the application decides what lines mean

One-Line Group(s) – for very little content: [USER]::Username|Skippy::PW|1234::[EOG]

• The :: directly after ] is the signal (a header is [NAME] then end-of-line); :: separates segments; must end ::[EOG]
• Data containing :: is backtick-quoted: [X]::Url|http://a::b::[EOG]

Key-Value Group(s) – [=SETTINGS=] promises every line is exactly one Key|Value pair; a value may be [{TEXT-GROUP}]

Line(s) – an array of fields on |: Protocol|RS232|9600!8!N!1!none

• The primary split is the format's; the nested split on ! is the application's, applied to a field on request; one level
• A leading or trailing | is an empty field at that end
• Per-line override: a line beginning with the slot-1 character then a one-shot primary delimiter: :!URL!https://x/?a=1|b=2

Escaping – \| and \! only; backslash is otherwise literal: \\server\share

• A field ending in \ takes a space before the delimiter: C:\Web\ |next – the space "escapes" the pair

Quoted strings – a field whose first non-space character is the quote character (backtick) runs to the quote character before the next delimiter or end-of-line

• Everything inside is literal – pipes, !, backslashes, spaces, ::; any other position is an ordinary character

Tables – first data line {id|name|email} names the fields; rows read into those positions; || empty; a row may stop early

• Appended fields name:::value – a field containing ::: is an appended field wherever it sits; positional fields end at the first one
• Kirk|Siqveland|AKA:::Cornelius Vogelgriff – no empty columns for everyone else, no truncation; tables only
• A positional value that contains ::: is backtick-quoted

Text-Group(s) – [{LICENSE}] – content as-is from the next line to the mandatory [EOG]; not closed by empty lines or headers

• No escaping, no trimming, no delimiters; full UTF-8; line endings kept; the newline before [EOG] is removed
• Referenced from a field as [{LICENSE}]; anywhere in the file; unreferenced = a long comment

[THIS-FILE] – reserved; optional; first Group when present; author notes about the whole file

• Author|… Created|… Description|… Generator|… – plus Trim|off, Encode|UTF-8, Localize|NFC|en-US|LTR

[EOF] – optional; at the start of a line; nothing after it is parsed

SetTags – a whole Set|File inside another file's comments: {SETTAG:NAME} … {/SETTAG:NAME/}, each on its own line

• One space after the host comment marker on every line; the opening marker stands as line 1 of the head
• In-group comments take two spaces after the marker; in block comments lines start at column one

X-Set – .xset; same core plus loaded modules; first intended module: Nested Groups (close by name, explicit [EOG]`, fields are the floor)

SET-File_OnePage_v4.5 – www.setfiles.org