The document discusses pure functions, which always return the same result for the same inputs, do not depend on hidden state or external inputs that could change, and do not mutate their arguments. Pure functions have advantages like being easy to test, reuse, reason about, and use in multithreading. Some examples of pure and impure functions are provided.

1. λ pure functions by Per Arneng 2015-02-08
pure
functions
pure
functions

2. λ pure functions by Per Arneng 2015-02-08
pure functions
● It always evaluates the same result value given
the same argument value(s)
● The result value cannot depend on any hidden
information or state that may change as
program execution proceeds or between
different executions of the program
● It can not depend on any external input from I/O
devices
pj f ŋk nsʊɚ ə ʃə

3. λ pure functions by Per Arneng 2015-02-08
pure functions
● It always evaluates the same result value given
the same argument value(s)
● The result value cannot depend on any hidden
information or state that may change as
program execution proceeds or between
different executions of the program
● It can not depend on any external input from I/O
devices
pj f ŋk nsʊɚ ə ʃə

4. λ pure functions by Per Arneng 2015-02-08
pure functions
● It always evaluates the same result value given
the same argument value(s)
● The result value cannot depend on any hidden
information or state that may change as
program execution proceeds or between
different executions of the program
● It can not depend on any external input from I/O
devices
pj f ŋk nsʊɚ ə ʃə

5. λ pure functions by Per Arneng 2015-02-08
ex:
int addOne(int value) {
return value + 1;
}

6. λ pure functions by Per Arneng 2015-02-08
impure:
void addOne(int *value) {
*value = *value = 1;
}

7. λ pure functions by Per Arneng 2015-02-08
impure:
void addOne(int *value) {
*value = *value = 1;
}
The argument is mutated

8. λ pure functions by Per Arneng 2015-02-08
impure 2:
int addOne(int value) {
int newValue = value + 1;
file.write(newValue);
return newValue;
}

9. λ pure functions by Per Arneng 2015-02-08
impure 2:
int addOne(int value) {
int newValue = value + 1;
file.write(newValue);
return newValue;
}
Could fail making the function
non deterministic and impure

10. λ pure functions by Per Arneng 2015-02-08
impure 2:
int addOne(int value) {
int newValue = value + 1;
file.write(newValue);
return newValue;
}
Using a hidden state

11. λ pure functions by Per Arneng 2015-02-08
advantages
● Easy to test
● Easy to reuse
● Easy to reason about (read, understand)
● Perfect for use in multithreading

12. λ pure functions by Per Arneng 2015-02-08
testability
● You do not need to mock hidden state resulting
in smaller and faster tests
● You only need to focus and arguments and
return values
● Tests that are not mocked are executed faster
than mocked ones. For hundreds of tests it will
matter

13. λ pure functions by Per Arneng 2015-02-08
maintainability
● Usually smaller blocks of code that can just be
copied out and reused in another context
● Easy to understand and reason about so you
are less likely to feel that you need to write your
own
● Easy to use as building blocks for other
functions and methods as inline or passed as
parameters

14. λ pure functions by Per Arneng 2015-02-08
readability
● When a pure function is used you know that it
will not modify anything making it easier to read
that code
● When you read a pure function your brain does
not have to keep external state in mind

15. λ pure functions by Per Arneng 2015-02-08
multithreading
● A pure function does not mutate it's argument
or make use of any hidden state so its easier to
use in a multithreaded context without having to
resort to synchronization

16. λ pure functions by Per Arneng 2015-02-08
more
● https://github.com/PerArneng/pfnomock
A repo showing a pure and non pure function
for handling transfers between accounts and
what that means in terms of testing with and
without mocking