Handling Null in Functional Programming

Nulls are a fact of life—sometimes a value can be absent, which is important information to have However if a variable can be null, we often have to write extra code to handle that possibility. How can we utilize functional, typed languages to only allow null where it communicates important information and eliminate it everywhere else? This talk will identify patterns to do just that and discuss how they can help shape your application’s architecture.

Given at Empex LA (http://empex.co/la.html)

Transcript

1. Handling Null Evelyn Masso (@outofambit)
3. What is “null”?
4. “not any” in latin
5. also known as nil, nothing, none
6. type nullable<T> = null | T option type
7. option type
8. type nullable<Number> = null | Number option type
9. function(amount: nullable<Number>) { if(amount !=="" null) { // do something
 } else { // ???? } } using an optional number
10. function(amount: nullable<Number>) { if(amount !=="" null) { // do something
 } else { // ???? } } using an optional number
11. Dropdown menu state
13. Peeta Gale
20. The tradeoff
21. When to use null • when something can optionally exist • instead of using a number (like -1) or a boolean type union • for data structures (but then check for it as soon as possible in business logic)
22. When to avoid null? • there are multiple ways something can not exist (that you care about) • future proofing your data model
24. Alternatives to null
25. enum People { Peeta, Gale, Hermione, } explicit cases enum Options { NoOne, SomeoneElse, Unknown }
26. type DatingOptions = People | Options explicit cases
27. default values enum Dates = { Peeta, Gail, Hermione, None } type Availables = Dates function(date: Availables = Dates.None) { // do something }
28. The tradeoff
30. Thanks! (@outofambit)