Fern

Opus #5

A clean and simple storage or configuration format with easy readability, minimal meta-characters, lenient parsing and simple conversion.

This is built to be a minimalist alternatives to popular configuration formats such as YAML or JSON, but without their individual pitfalls, such as YAML's strict whitespace grammar and JSON's reliance on quotes and brackets.

Fern is also designed to be easily customisable and extendable, allowing a user to build domain-specific features into the general parser or framework to build handlers for custom data types.

Fern ignores all whitespace, except for the single character that separates entities.

Data Types

Type	Delimiter	Description
Branch	()	A new sub-branch, a key value map.
List	[]	A list.
String	"text"	Simple strings, supports both simple character escapes and multi-line content.
Integer	100	Simple int-32.
Long	100L	Simple int-64. Suffix is only required to differentiate from an integer in small numbers.
Short	100S	A short. Suffix is always required.
Double	100.0D	A double. Suffix is not required, unless the number is not a decimal.
Float	0.5F	A floating point number. Suffix is always required.
Byte	13B	A byte. Suffix is always required.
Null	null	A null-value.
Boolean	true	A boolean true/false value.
Insight	<Type>	A type insight provided by a third-party library.

Domain-specific data types can be added by custom parser implementations by adding a new ValueHandler - they require only a recognisable identifier.
An example would be a Bird class. A parser can register <Bird> as a special identifier, which would pass the subsequent value to the desired value handler.

Comments

Some fern implementations support a style of comment marked by a ` backtick character.

These comments are an entity like keys and values and so require some separation.

key `comment` value
key value `comment`
`this is a
really long comment`

Examples

Basic uses of types.

name "Tony"
age 42    
height 1.85 
weight 70.2 
spouse null  
likes [ "chicken" "pasta" "bacon" ]
dislikes [ "fish" "ice cream" 74 ]
education (   
    school "Hartlepool Secondary School"
    college "Wisham College"
    university null
)

More examples of type usage.

a_key ( 
    abc 1 
    xyz 2D 
    pqr 5.0 
)
another_key "value" 

Fern does not require whitespace, except to separate entities. This means a user can mix indentation units to their heart's content.

map ( key "value" thing 66 ) list [ 1 2 3 ]

Maven Information

<repository>
    <id>kenzie</id>
    <name>Kenzie's Repository</name>
    <url>https://repo.kenzie.mx/releases</url>
</repository>

<dependency>
    <groupId>mx.kenzie</groupId>
    <artifactId>fern</artifactId>
    <version>2.0.0</version>
</dependency>

Binary Data

The Java implementation of Fern contains a binary data format for more compressed values. Naturally, using this eliminates Fern's potential as an editable configuration format.

The binary format is supplied as a simple way to compress the internal data. The BinaryFern object is fairly interchangeable with the regular Fern but has no support for third-party handlers: only the built-in data types are supported.

Adding Types

Fern is built to support additional value types for domain-specific use.

A type must have a recognisable starting character (e.g. .) and a recognisable end. Types are responsible for recognising their own end, so this may not be a single distinct character.