Skip to main content

Data Types

Numbers​

R7RS specification define Numerical Tower, different type of numbers and operation on them. But it's not required for Scheme implementation to add support for them. Some Scheme implementation do support it.

Basic numbers are:

  • Integers, e.g. 10, 20, or 1000
  • floats, e.g. 1.2, 1e-2 (you can use scientific notation)
  • rationals e.g. 1/2, 3/4, or 1/10
  • complex numbers 10+10i, 1/2+1/2i

There is also notion of exactness in those numbers. Inexact numbers are floats, the rest of the numbers are exact because they don't give any rounding errors like floats do (this is how IEEE 754 standard for floating numbers work).

There are also special constant +nan.0, -nan.0, +inf.0, and -inf.0. Positive and negative Not a number object and negative and positive infinity.

You can convert string to numbers with string->number procedure and number to string with number->string.

(string->number "1001" 2)
;; ==> 9
(string->number "1001")
;; ==> 1001

The first argument is a string and the second one is the base of the number. In similar way you can use number->string:

(number->string 1001 2)
;; ==> "1111101001"
(number->string 1001 16)
;; ==> "3e9"

Boolean values​

Scheme define two boolean constants #f and #t but note that the only false value, according to R7RS specification, should be #f. The specification also defines #true and #false aliases. Some Scheme also defines true and false without hash.

Strings​

Strings in Scheme use only double quote symbols. They can be multiline. If you want to add double quote, you need to escape with the slash "this is \"Scheme\" language".

You can also inject hex representation of a character with "\x1B;" this will create a string with Escape character used by Terminal emulators to add formatting (like colors).

Characters​

You can define single character as data type

#\A
#\B
#\c

you can use characters to form a string:

(string #\h #\e #\l #\l #\o)

This evaluates into string "hello"

You can also split the string into individual characters:

(string->vector "hello")

This evaluates into vector of characters: #(#\h #\e #\l #\l #\o)

Symbols​

Symbols are special type of objects that are similar to string but without quotes. They can appear as variable names but can also be used as values, but this requires quotation.

Valid symbols:

foo
bar
baz

You can also convert string to symbol with:

(string->symbol "hello")

Scheme have conversion of using -> arrow to define procedure that convert types.

In R7RS spec there is also a way to define symbols with special characters. Those symbols start and end with a vertical bar character.

(define |foo bar| 10)
(define |foo () bar| 20)
(display |foo bar|)
;; ==> 10
(display |foo () bar|)
;; ==> 20

You can also put escaped characters in symbol, like in strings:

(symbol=? '|\x3BB;| 'λ)
;; ==> #t

Comments​

There are 3 types of comments in Scheme:

  1. ;semicolon create comments to the end of the line
  2. #; quote single S-Expression (list/tree or atoms)
  3. #| |# those are multiline comments that can wrap any text inside

Empty list​

An empty list is a special object in Scheme that indicates end of the list.

()

Pairs​

The base of Lisp and Scheme are pairs also called cons cells. You can create them with cons operation:

(cons 1 2)

or with dot syntax:

(1 . 2)

But the second example requires quotation otherwise Scheme will try to evaluate 1 as a function.

If cons cells are put into e sequence:

(1 . (2 . (3 . ())))

The last cell is a pair of 3 and empty list. If you create cons like this, it's simplified as a list when printed:

(1 2 3)

To create the same list, you can use list procedure:

(list 1 2 3)

To get the first element of the list you use procedure car and to get the rest of the list you use cdr. So car and cdr returns first and second element of the Pair (cons cell).

Scheme used car and cdr for historical reasons. The first lisp interpreter was using address registers of IBM 704 mainframe computer.

  • car stands for Contents of the Address part of Register
  • cdr stands for Contents of the Decrement part of Register

Scheme should also define abbreviations for list accessors:

Example caddr is the third element of the list. It's the same as (car (cdr (cdr x))). Often, Scheme and lisp interpreters define up to 5 combinations of d and a to get different elements out of a list.

List Modification​

You can modify the lists (cons cells) using two functions:

  • set-car!
  • set-cdr!

By convention, if a procedure modifies something it's indicated with an exclamation mark.

Improper list​

Proper list is a list with each cdr be a list or empty list at the end:

'(1 2 3 4 5)

in the other hand, if you have a list that end with something else then an empty list, you have not a valid list. This list is often called improper list. It looks like this:

'(1 2 3 . 4)

Cycles​

You can create list cycles directly when defining your data structure with datum syntax.

It looks like this:

#0=(1 2 . #0#)

This will create an infinite list of values (1 2 1 2 1 2 ...). #0 indicate a pointer and #0# a reference to the beginning of the list.

Vectors​

Vectors are created like list, but they have hash in front:

#(1 2 3 4)

This will create an immutable vector that can't be changed. To create a vector that can be modified you can use:

(vector 1 2 3)

Quotations​

By default if you write lists they are treated as code. To create a data you need quotations.

Base Quote​

To create basic quotation you use single quote character:

'(1 2 3 4)

When Scheme find this expression it will not try to execute the function 1 only return this list as data.

Vector syntax is automatically quoted.

'#(1 2 3)

is the same as:

#(1 2 3)

Quasiquote​

There is also different type of quotation that allow to execute part of the expression. It's called quasi quote. To create quasi quote you need back tick symbol. That's why it's often called back quotation.

`(1 2 ,(+ 2 1))

A comma is special syntax that can be used inside quasi quote to indicate that the expression after it should be evaluated and inserted into the list. This will evaluate into the same expression as:

'(1 2 3)

There is also another escape symbol which is ,@. It works similar to a comma, but the data inside (it must be a list) is spliced into the outer list.

`(1 2 3 ,@(list 4 5 6))

The result expression will look like this:

'(1 2 3 4 5 6)

Quotation of quotation​

If you quote the quotation, you will get expression that change special symbols with S-Expressions:

''(1 2 3)

Will return:

(quote (1 2 3))

Because comma is just an alias for quote. You can use them interchangeably. But using symbols is faster to type. If you quote quasi quote expression, you will also get symbols expanded:

'`(1 2 3 ,(+ 1 2) ,@(list 4 5))

This will be the output:

(quasiquote (1 2 3 (unquote (+ 1 2)) (unquote-splicing (list 4 5))))

Quotation on vector literals​

Quasiquote also works with vector literals.

`#(1 2 ,(+ 1 2))
;; ==> #(1 2 3)
`#(1 2 ,@(list 3 4))
;; ==> #(1 2 3 4)

Special symbols​

Most special symbols in Scheme start with a hash symbol. Example are Byte vectors

#u8(1 2 3 4)

Above creates 8 bit byte vector of numbers. In R7RS, only unsigned 8 bit vectors are defined. But in SRFI-4 are more bit vectors types. They all starts with hash. In different SRFI there are more examples of syntax's that start with hash. This is just a convention everything is using.