> For the complete documentation index, see [llms.txt](https://docs.internetobject.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.internetobject.org/appendices/grammar.md).

# Formal Grammar (EBNF)

The grammar of Internet Object in EBNF.

This appendix gives the grammar of Internet Object in EBNF. It is the normative reference for document structure; the prose chapters explain semantics. Lexical rules (whitespace, comments, literals) are listed at the end.

> Notation: `=` defines a rule, `|` alternation, `[ ]` optional, `{ }` zero-or-more, `( )` grouping, `"…"` a literal. Whitespace and comments may appear between tokens unless a rule states otherwise.

## Document

```ebnf
document        = [ header "---" ] section { sectionBreak section }
                | value ;                         (* a bare, header-less value/object *)

header          = { definition } ;
section         = [ sectionTag ] ( collection | object ) ;
sectionBreak    = "---" [ sectionTag ] ;
sectionTag      = sectionName [ ":" schemaRef ]    (* e.g.  employee : $employee *)
                | schemaRef ;                      (* e.g.  $employee *)
```

## Header definitions

```ebnf
definition      = "~" key ":" defValue ;
key             = metaKey | variableKey | refKey ;
metaKey         = name ;                            (* metadata *)
variableKey     = "@" name ;                        (* value variable *)
refKey          = "$" name ;                        (* schema/type reference *)
defValue        = value | schema ;
schemaRef       = "$" name ;
```

## Collections and records

```ebnf
collection      = record { record } ;
record          = "~" [ recordBody ] ;             (* "~" alone = empty object {} *)
recordBody      = object | openObject | value ;    (* a bare value is promoted to an object *)
```

## Objects and arrays

```ebnf
object          = "{" [ memberList ] "}" ;
openObject      = memberList ;                      (* unbraced, top-level only *)
memberList      = member { "," member } [ "," ] ;
member          = [ key ":" ] [ value ] ;          (* keyed or positional; value may be absent *)

array           = "[" [ valueList ] "]" ;
valueList       = value { "," value } ;
```

> **A member's value may be absent**, which is how an **empty position** is written: `{p, , q}` holds three members, the middle one carrying no value. A writer is [required to emit this form](/serialization/document-output.md#absent-members-hold-their-place) when a positional member is skipped, so a grammar that demanded a value here would reject conformant writer output.
>
> An array has no such production: `valueList` requires a value between commas, so `[a,,c]` is **not** an array with a hole — it is an error (`unexpected-token`). Objects bind members by position and therefore need to express a gap; arrays do not.
>
> The optional trailing comma in `memberList` closes the list; it does **not** add an empty final position. `{a, b,}` holds **two** members, not three.

## Values

```ebnf
value           = object | array | scalar | variableRef ;
variableRef     = "@" name | "$" name ;
scalar          = string | number | specialNumber | bigint | decimal
                | datetime | binary | boolean | null ;
```

## Schema (MemberDef and SchemaDef)

```ebnf
schema          = schemaDef | memberDef | refKey ;
schemaDef       = "{" memberDefList "}" | memberDefList ;   (* object shape *)
memberDefList   = schemaMember { "," schemaMember } [ "," "*" [ ":" memberDef ] ] ;
                                                    (* the wildcard `*` is BARE; a quoted "*"
                                                       is an ordinary memberName *)
schemaMember    = memberName [ ":" ( type | memberDef | array | schemaRef ) ] ;
memberName      = name [ "?" ] [ "*" ]              (* suffixes belong to the bare-name token *)
                | regularString | rawString ;       (* quoted: literal, no suffixes -- use
                                                       optional: / "null": in the memberDef *)
memberDef       = "{" type { "," option } "}" ;            (* type + constraints *)
option          = positionalValue | ( name ":" value ) ;   (* e.g. min: 0 *)
type            = "string" | "int" | "uint8" | "bool" | "datetime" | "decimal" | … ;
```

## Lexical

```ebnf
string          = openString | regularString | rawString ;
regularString   = '"' { char | escape } '"' | "'" { char | escape } "'" ;
rawString       = "r" ( '"' { rawChar } '"' | "'" { rawChar } "'" ) ;
openString      = unquotedChar { unquotedChar } ;          (* ends at a structural char *)

number          = [ sign ] ( decInt [ frac ] [ exp ]    (* frac/exp: base-10 only *)
                           | frac [ exp ]               (* leading-dot form, e.g. .5 *)
                           | hex | octal | binaryNum ) ;
hex             = ( "0x" | "0X" ) hexDigit { hexDigit } ;
octal           = ( "0o" | "0O" ) octDigit { octDigit } ;
binaryNum       = ( "0b" | "0B" ) ( "0" | "1" ) { "0" | "1" } ;
bigint          = [ sign ] ( decInt | hex | octal | binaryNum ) "n" ;
decimal         = [ sign ] ( decInt [ frac ] | frac ) "m" ;

datetime        = "dt" quoted | "d" quoted | "t" quoted ;  (* ISO-8601 inside quotes *)
binary          = "b" quoted ;                              (* base64 inside quotes *)

boolean         = "T" | "F" | "true" | "false" ;
null            = "N" | "null" ;
specialNumber   = "NaN" | "Inf" | "+Inf" | "-Inf" ;   (* reachable via `scalar` *)

comment         = "#" { anyCharExceptNewline } ;
ws              = ? Unicode whitespace ? ;
name            = nameStart { nameChar } ;
sectionName     = sectionChar { sectionChar } ;    (* ANCHORED: the whole name must match *)
sectionChar     = letter | mark | digit | "-" | "_" ;
```

> `sectionName` is narrower than `name`, and it is the one name in the format with **no quoted form** — the separator line runs to the end of the line, so nothing bounds the name. The rule is anchored on purpose: a parser must not match a prefix and leave the remainder to fail later, or `--- a,b: $x` would silently become a section named `a`. See [Data Sections](/structure-and-syntax/introduction/data.md#section-names-are-bare-names).

> This grammar is a working draft for the 1.0 specification. Edge cases (precise open-string termination, escape sequences, and datetime sub-formats) are described in their respective chapters and will be folded in as the grammar is finalized.

## See Also

* [Structural Elements](/structure-and-syntax/structural-elements.md)
* [Value Representations](/structure-and-syntax/values.md)
* [Internet Object Schema](/schema-definition-language/internet-object-schema.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.internetobject.org/appendices/grammar.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.