Introduction: Why SML
Before you start writing the first line of SML, answer a question: * * Why is there another configuration format? **
JSON, YAML, and TOML are all very useful, but they each have their own frustrating aspects:
| Format | Pain Points |
|---|---|
| JSON | Quotation marks, commas, and curly braces must not be missing at all; Writing configuration is very verbose |
| YAML | Indent sensitive, one space misplacement will explode; no will be mistaken for false, which will startle you |
| TOML | If nested deeply, many [table.sub] headers need to be written, and the block structure is not intuitive |
Design Objectives of SML
SML wants to do “Write configurations like notes”:
-
Less ceremonial sense - quotation marks can be selected, commas can be selected, and colon blocks can be omitted.
-
Structure relies on curly braces - without indentation, copying and pasting will not result in strange errors.
-
Prioritize readability - making it easy to understand at a glance and making diff/review easier.
-
Customizable - Basic capabilities are minimal, while complex capabilities (including, regular, etc.) are enabled as needed, avoiding the pitfalls of excessive YAML complexity.
-
Built in contract - not only “storing data”, but also defining “what data should look like” and verifying it during the parsing period.
A minimum comparison
The same configuration can be written in three ways:
{
"name": "gateway",
"port": 8080,
"debug": true,
"tags": ["logging", "metrics"]
}
name: gateway
port: 8080
debug: true
tags:
- logging
- metrics
name: gateway
port: 8080
debug: true
tags: [ logging metrics ]
The SML version has almost no punctuation noise - no quotes, no commas, no indentation traps. This is its core temperament.
What is it not suitable for
SML is a data/configuration format, not a programming language:
-Cannot write if/for, cannot define functions, cannot calculate 1+1.
-When logic is needed, read SML in the host language (Rust/JS/Lua…) and process it in the code.
Remember one sentence: * * SML is responsible for “describing what it is”, and code is responsible for “how to do it”. **
In the next chapter, we will write the first real SML file. → Chapter 1
Hands on practice
After reading this chapter, directly modify SML in the editor below and click “Run” to immediately see the parsing results or validation errors - having output is necessary for efficient learning.