Syntax Reference

Syntax Reference

LOVE-LANG has an intentionally minimal surface syntax. It is formally defined by an ABNF grammar derived from “Forma — A Formal Naturalized Language.”

---

Documents

A LOVE document is a series of base statements, whitespace, and comments:

forma-doc = *(ws / comment / base-statement)

---

Base Statements

The fundamental unit of LOVE code is the base statement — assigning a value to a named key:

key = value
key is value    # 'is' is an alias for '='

base-statement = key-def ws assign-op ws value ws *(newline)
assign-op      = "=" / "is"

Examples

statements.love

name        = "LOVE-LANG"
version     = 0.0.1
active      = true
description is "A linguistics-oriented programming language"

---

Type Annotations

You can optionally annotate the type of any key using two syntaxes:

name @[String]   = "Alice"    # angle-bracket style
age  :Number     = 30         # colon style

type-annotation = ( "@" "[" identifier "]" )
                / ( ":" identifier )

---

Atoms (Primitive Values)

Atom type	Example	Description
string	"hello world"	Double-quoted UTF-8 string
number	42, 3.14	Integer or float
boolean	true, false	Boolean literal
symbol	:keyword	Lisp-style named constant
identifier	myValue	Bareword reference
datetime	(ISO 8601)	Date/time value

---

The Five Expression Levels

LOVE uses bracket depth to encode semantic complexity and trust. This is not merely syntactic sugar — each level has a distinct meaning in the evaluator’s topology.

L1 — Data Expressions [x]

Simple S-expressions. Variables and primitive references.

colours = [red green blue]
pair    = [1 2]
empty   = []

L2 — Pure Operations [[x]]

Standard library math and pure functions. No side-effects.

sum     = [[ + 1 2 3 ]]
product = [[ * 6 7 ]]
max-val = [[ Math.max [1 9 3 7] ]]

L3 — Type Contracts [[[x]]]

Structural schemas and type validators. Used for type checking.

UserPayload = [[[
  name   @[String]
  email  @[String]
  age    @[Number]
]]]

L4 — Compiler Directives [[[[x]]]]

Macros and compile-time transformations. Evaluated before the programme runs.

config = [[[[
  #mode    Production
  #target  wasm
  #inline  true
]]]]

L5 — System FFI / I/O [[[[[x]]]]]

Side-effecting operations: file I/O, network, database writes. The highest trust level.

[[[[[
  @sql_write "users" { :name "Alice" :age 30 }
]]]]]

[[[[[
  @http_post "https://api.example.com/record" payload
]]]]]

Caution

Level 5 expressions are restricted by the Superautology. They can only use effectors that are explicitly declared in model.edn. Undefined effectors cause a hard compiler error.

---

Forms

The “stuff inside” brackets — forms are recursive and data-only:

form = ws / comment / atom / expression

---

Comments

Comments begin with # and extend to end of line:

# This is a comment
name = "LOVE"   # inline comment

---

Identifiers and Symbols

# Valid identifiers
myValue
my-value
my_value?
my-predicate!

# Symbols (Lisp-style keywords)
:name
:type
:active

identifier = 1*(ALPHA / "_") *ID-CHAR
ID-CHAR    = ALPHA / DIGIT / "-" / "_" / "?" / "!"
symbol     = ":" 1*ID-CHAR

---

Complete ABNF Reference

The formal grammar is reproduced here in full from syntax.abnf:

syntax.abnf

; 1. TOP LEVEL
forma-doc = *(ws / comment / base-statement)

; 2. BASE STATEMENT
base-statement = key-def ws assign-op ws value ws *(newline)
key-def        = identifier [ws type-annotation]
assign-op      = "=" / "is"
value          = atom / expression

; 3. TYPE ANNOTATION
type-annotation = ( "@" "[" identifier "]" )
                / ( ":" identifier )

; 4. CORE EXPRESSIONS
expression = meta-expr-5 / meta-expr-4 / macro-expr / type-expr / data-expr
data-expr  = "["     *form "]"
type-expr  = "[["    *form "]]"
macro-expr = "[[["   *form "]]]"
meta-expr-4 = "[[[[" *form "]]]]"
meta-expr-5 = "[[[[["*form "]]]]]"

; 5. FORM
form = ws / comment / atom / expression

; 6. ATOMS
atom = symbol / identifier / string / number / boolean / datetime
symbol     = ":" 1*ID-CHAR
identifier = 1*(ALPHA / "_") *ID-CHAR
ID-CHAR    = ALPHA / DIGIT / "-" / "_" / "?" / "!"

; 7. WHITESPACE & COMMENTS
ws      = 1*(%x20 / %x09)
comment = "#" *(%x20-%x10FFFF) %x0A
newline = %x0A