Domain Driven Design with the F# type System -- F#unctional Londoners 2014

Statically typed functional programming languages like F# encourage a very different way of thinking about types. The type system is your friend, not an annoyance, and can be used in many ways that might not be familiar to OO programmers.

Types can be used to represent the domain in a fine-grained, self documenting way. And in many cases, types can even be used to encode business rules so that you literally cannot create incorrect code. You can then use the static type checking almost as an instant unit test — making sure that your code is correct at compile time.

In this talk, we'll look at some of the ways you can use types as part of a domain driven design process, with some simple real world examples in F#. No jargon, no maths, and no prior F# experience necessary.

Code, links to video, etc., at http://fsharpforfunandprofit.com/ddd

NEW AND IMPROVED - added sections on:
* why OO, not FP is scary
* designing with states and transitions

Transcript

1. type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool } // true if ownership of // email address is confirmed Domain Driven Design with the F# type system Prologue: how many things are wrong?
2. type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool } Prologue: which values are optional?
3. type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool } Prologue: what are the constraints?
4. type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool } Prologue: what groups are atomic?
5. type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool } Prologue: domain logic?
6. Prologue: F# can help type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool }
7. Domain Driven Design with the F# type system Scott Wlaschin @ScottWlaschin fsharpforfunandprofit.com FPbridge.co.uk /ddd
8. What is DDD?
9. Functional Programming Domain Modelling
10. What I’m going talk about: • Demystifying functional programming • Functional programming for real world applications • F# vs. C# for domain driven design • Understanding the F# type system • Designing with types
11. Demystifying functional programming Why is it so hard?
12. Functional programming is scary
13. Functional programming is scary
14. Object oriented programming is scary
15. Functional programming is scary
16. Functional programming for real world applications
17. Functional programming is... ... good for mathematical and scientific tasks ... good for complicated algorithms ... really good for parallel processing ... but you need a PhD in computer science 
18. Functional programming is good for... Boring Line Of Business Applications (BLOBAs)
19. Must haves for BLOBA development... • Express requirements clearly • Rapid development cycle • High quality deliverables • Fun
21. F# vs. C# for Domain Driven Design A simple immutable object
22. How do you implement a Value object? Equality based on comparing all properties PersonalName: FirstName = "Alice" LastName = "Adams" PersonalName: FirstName = "Alice" LastName = "Adams" Equal  Therefore must be immutable
23. class PersonalName { public PersonalName(string firstName, string lastName) { this.FirstName = firstName; this.LastName = lastName; } public string FirstName { get; private set; } public string LastName { get; private set; } } Value object definition in C#
24. class PersonalName { public PersonalName(string firstName, string lastName) { this.FirstName = firstName; this.LastName = lastName; } public string FirstName { get; private set; } public string LastName { get; private set; } } Value object definition in C#
25. class PersonalName { // all the code from above, plus... public override int GetHashCode() { return this.FirstName.GetHashCode() + this.LastName.GetHashCode(); } public override bool Equals(object other) { return Equals(other as PersonalName); } public bool Equals(PersonalName other) { if ((object) other == null) { return false; } return FirstName == other.FirstName && LastName == other.LastName; } Value object definition in C# (extra code for equality)
26. type PersonalName = {FirstName:string; LastName:string} Value object definition in F#
27. This page intentionally left blank Value object definition in F# (extra code for equality)
28. How do you implement an Entity object? Equality based on some sort of id Person: Id = 1 Name = "Alice Adams" Person: Id = 1 Name = "Bilbo Baggins" Equal X  Generally has mutable content
29. class Person { public Person(int id, PersonalName name) { this.Id = id; this.Name = name; } public int Id { get; private set; } public PersonalName Name { get; set; } } Entity object definition in C# (part 1)
30. class Person { // all the code from above, plus... public override int GetHashCode() { return this.Id.GetHashCode(); } public override bool Equals(object other) { return Equals(other as Person); } public bool Equals(Person other) { if ((object) other == null) { return false; } return Id == other.Id; } } Entity object definition in C# (part 2)
31. [<CustomEquality; NoComparison>] type Person = {Id:int; Name:PersonalName} with override this.GetHashCode() = hash this.Id override this.Equals(other) = match other with | :? Person as p -> (this.Id = p.Id) | _ -> false Entity object definition in F# with equality override
32. Entity object definition in F# with no equality allowed [<CustomEquality; NoComparison>] type Person = {Id:int; Name:PersonalName} [<NoEquality; NoComparison>]
33. type Person = { ... ... ... } let tryCreatePerson name = // validate on construction // if input is valid return something // if input is not valid return error   Advantages of immutability
34. class PersonalName { public PersonalName(string firstName, string lastName) { this.FirstName = firstName; this.LastName = lastName; } public string FirstName { get; private set; } public string LastName { get; private set; } public override int GetHashCode() { return this.FirstName.GetHashCode() + this.LastName.GetHashCode(); } public override bool Equals(object other) { return Equals(other as PersonalName); } public bool Equals(PersonalName other) { if ((object) other == null) { return false; } return FirstName == other.FirstName && LastName == other.LastName; } } Reviewing the C# code so far... class Person { public Person(int id, PersonalName name) { this.Id = id; this.Name = name; } public int Id { get; private set; } public PersonalName Name { get; set; } public override int GetHashCode() { return this.Id.GetHashCode(); } public override bool Equals(object other) { return Equals(other as Person); } public bool Equals(Person other) { if ((object) other == null) { return false; } return Id == other.Id; } } : IValue : IEntity
35. [<StructuralEquality;NoComparison>] type PersonalName = { FirstName : string; LastName : string } Reviewing the F# code so far... [<NoEquality; NoComparison>] type Person = { Id : int; Name : PersonalName }
36. Comparing C# vs. F# C# F# Value objects? Non-trivial Easy Entity objects? Non-trivial Easy Value objects by default? No Yes Immutable objects by default? No Yes Can you tell Value objects from Entities at a glance? No Yes Understandable by non-programmer? No Yes C# vs. F# for DDD
37. F# for Domain Driven Design Communicating a domain model
38. Communication is hard... U-N-I-O-N-I-Z-E U-N-I-O-N-I-Z-E U-N-I-O-N-I-Z-E
39. Communication in DDD: “Bounded Context” Business Chemistry un-ionize unionize Supermarket Email System Spam Spam
40. Communication in DDD: “Bounded Context” Business Chemistry un-ionize unionize Supermarket Email System Spam Spam Sales Warehouse Product Product
41. Communication in DDD: “Bounded Context” Business Chemistry un-ionize unionize Supermarket Email System Spam Spam Sales Warehouse Product Product Marketing Finance Customer Customer
42. Communication in DDD: “Ubiquitous Language” Chemistry Ion Atom Molecule Polymer Compound Bond
43. Communication in DDD: “Ubiquitous Language” Chemistry Ion Atom Molecule Polymer Compound Bond Sales Product Promotion Customer Tracking
44. Communication in DDD: “Ubiquitous Language” Chemistry Ion Atom Molecule Polymer Compound Bond Sales Product Promotion Customer Tracking Warehouse Product Stock Transfer Depot Tracking
45. module CardGame = type Suit = Club | Diamond | Spade | Heart type Rank = Two | Three | Four | Five | Six | Seven | Eight | Nine | Ten | Jack | Queen | King | Ace type Card = Suit * Rank type Hand = Card list type Deck = Card list type Player = {Name:string; Hand:Hand} type Game = {Deck:Deck; Players: Player list} type Deal = Deck –› (Deck * Card) type PickupCard = (Hand * Card) –› Hand '*' means a pair. Choose one from each type Ubiquitous language list type is built in
46. module CardGame = type Suit = Club | Diamond | Spade | Heart type Rank = Two | Three | Four | Five | Six | Seven | Eight | Nine | Ten | Jack | Queen | King | Ace type Card = Suit * Rank type Hand = Card list type Deck = Card list type Player = {Name:string; Hand:Hand} type Game = {Deck:Deck; Players: Player list} type Deal = Deck –› (Deck * Card) type PickupCard = (Hand * Card) –› Hand
47. module CardGame = type Suit = Club | Diamond | Spade | Heart type Rank = Two | Three | Four | Five | Six | Seven | Eight | Nine | Ten | Jack | Queen | King | Ace type Card = Suit * Rank type Hand = Card list type Deck = Card list type Player = {Name:string; Hand:Hand} type Game = {Deck:Deck; Players: Player list} type Deal = Deck –› (Deck * Card) type PickupCard = (Hand * Card) –› Hand
49. Understanding the F# type system An introduction to “algebraic” types An introduction to “algebraic” types
50. Composable types
51. Creating new types New types are constructed by combining other types using two basic operations: type typeW = typeX "times" typeY type typeZ = typeX "plus" typeY
52. Creating new types (a function) AddOne AddOne int –› int 1 2 3 4 2 3 4 5
53. Representing pairs AddPair ? –› int (1,2) (2,3) (3,4) (4,5) 3 5 7 9
54. Representing pairs × = (1,2) (2,3) (3,4) (4,5) 1 2 3 4 2 3 4 5
55. Representing pairs × = (true, false) (true, true) (false, false) (false, true) true false true false
56. Representing pairs pair of ints written int * int pair of bools written bool * bool
57. Using tuples for data × = Alice, Jan 12th Bob, Feb 2nd Carol, Mar 3rd Set of people Set of dates
58. Using tuples for data × = Alice, Jan 12th Bob, Feb 2nd Carol, Mar 3rd Set of people Set of dates type Birthday = Person * Date
59. Representing a choice Temp F IsFever ? –› bool true false Temp C or
60. Representing a choice + = 98˚ F 99˚ F 100˚ F 101˚ F 37.0˚ C 37.5˚ C 38.0˚ C 38.5˚ C Temp F Temp C or
61. Representing a choice + = 98˚ F 99˚ F 100˚ F 101˚ F 37.0˚ C 37.5˚ C 38.0˚ C 38.5˚ C Temp F Temp C or Tag these with “C” type Temp = | F of int | C of float
62. Using choices for data type PaymentMethod = | Cash | Cheque of int | Card of CardType * CardNumber
63. Working with a choice type type PaymentMethod = | Cash | Cheque of int | Card of CardType * CardNumber let printPayment method = match method with | Cash –› printfn “Paid in cash" | Cheque checkNo –› printfn “Paid by cheque: %i" checkNo | Card (cardType,cardNo) –› printfn “Paid with %A %A" cardType cardNo
64. Using choices vs. inheritance interface IPaymentMethod {..} class Cash : IPaymentMethod {..} class Cheque : IPaymentMethod {..} class Card : IPaymentMethod {..} type PaymentMethod = | Cash | Cheque of int | Card of CardType * CardNumber class Evil : IPaymentMethod {..} Data and code is scattered around many locations What goes in here? What is the common behaviour? OO version:
65. Summary: What are types for in FP? An annotation to a value for type checking type AddOne: int –› int Domain modelling tool type Deal = Deck –› (Deck * Card)
66. TYPE ALL THE THINGS
67. Designing with types What can we do with this type system?
68. Required vs. Optional type PersonalName = { FirstName: string; MiddleInitial: string; LastName: string; } required required optional
69. Null is not the same as “optional” Length string –› int “a” “b” “c” 1 2 3 “a” “b” “c” null
70. Spock, set phasers to null! That is illogical, Captain
71. Null is not the same as “optional” Length string –› int “a” “b” “c” null 1 2 3
73. Null is not allowed Length string –› int “a” “b” “c” null 1 2 3 X
74. A better way for optional values + = “a” “b” “c” “a” “b” “c” missing or Tag with “Nothing” type OptionalString = | SomeString of string | Nothing
75. type OptionalInt = | SomeInt of int | Nothing type OptionalString = | SomeString of string | Nothing type OptionalBool = | SomeBool of bool | Nothing Defining optional types
76. The built-in “Option” type type PersonalName = { FirstName: string MiddleInitial: string LastName: string } type Option<'T> = | Some of 'T | None
77. The built-in “Option” type type PersonalName = { FirstName: string MiddleInitial: Option<string> LastName: string } type Option<'T> = | Some of 'T | None
78. The built-in “Option” type type PersonalName = { FirstName: string MiddleInitial: string option LastName: string } type Option<'T> = | Some of 'T | None
79. Single choice types type Something = | ChoiceA of A type Email = | Email of string type CustomerId = | CustomerId of int
80. Wrapping primitive types Is an EmailAddress just a string? Is a CustomerId just a int? Use single choice types to keep them distinct type EmailAddress = EmailAddress of string type PhoneNumber = PhoneNumber of string type CustomerId = CustomerId of int type OrderId = OrderId of int
81. Creating the EmailAddress type let createEmailAddress (s:string) = if Regex.IsMatch(s,@"^\S+@\S+\.\S+$") then (EmailAddress s) else ? let createEmailAddress (s:string) = if Regex.IsMatch(s,@"^\S+@\S+\.\S+$") then Some (EmailAddress s) else None createEmailAddress: string –› EmailAddress createEmailAddress: string –› EmailAddress option
82. Constrained strings type String50 = String50 of string let createString50 (s:string) = if s.Length <= 50 then Some (String50 s) else None createString50 : string –› String50 option
83. Constrained numbers + – 999999 Qty: Add To Cart
84. Constrained numbers type OrderLineQty = OrderLineQty of int let createOrderLineQty qty = if qty >0 && qty <= 99 then Some (OrderLineQty qty) else None createOrderLineQty: int –› OrderLineQty option
85. type Contact = { FirstName: string MiddleInitial: string LastName: string EmailAddress: string IsEmailVerified: bool } The challenge, revisited
86. The challenge, revisited type Contact = { FirstName: string MiddleInitial: string option LastName: string EmailAddress: string IsEmailVerified: bool }
87. The challenge, revisited type Contact = { FirstName: String50 MiddleInitial: String1 option LastName: String50 EmailAddress: EmailAddress IsEmailVerified: bool }
89. type Contact = { Name: PersonalName Email: EmailContactInfo } The challenge, revisited type PersonalName = { FirstName: String50 MiddleInitial: String1 option LastName: String50 } type EmailContactInfo = { EmailAddress: EmailAddress IsEmailVerified: bool }
90. Encoding domain logic Rule 1: If the email is changed, the verified flag must be reset to false. Rule 2: The verified flag can only be set by a special verification service type EmailContactInfo = { EmailAddress: EmailAddress IsEmailVerified: bool }
91. Encoding domain logic type VerifiedEmail = VerifiedEmail of EmailAddress type EmailContactInfo = | Unverified of EmailAddress | Verified of VerifiedEmail type VerificationService = (EmailAddress * VerificationHash) –› VerifiedEmail option
92. type EmailAddress = ... type VerifiedEmail = VerifiedEmail of EmailAddress type EmailContactInfo = | Unverified of EmailAddress | Verified of VerifiedEmail The challenge, completed type PersonalName = { FirstName: String50 MiddleInitial: String1 option LastName: String50 } type Contact = { Name: PersonalName Email: EmailContactInfo }
93. Making illegal states unrepresentable type Contact = { Name: Name Email: EmailContactInfo Address: PostalContactInfo }
94. Making illegal states unrepresentable type Contact = { Name: Name Email: EmailContactInfo Address: PostalContactInfo } New rule: “A contact must have an email or a postal address”
95. Making illegal states unrepresentable type Contact = { Name: Name Email: EmailContactInfo option Address: PostalContactInfo option } New rule: “A contact must have an email or a postal address”
96. Making illegal states unrepresentable “A contact must have an email or a postal address” implies: • email address only, or • postal address only, or • both email address and postal address
97. Making illegal states unrepresentable type ContactInfo = | EmailOnly of EmailContactInfo | AddrOnly of PostalContactInfo | EmailAndAddr of EmailContactInfo * PostalContactInfo type Contact = { Name: Name ContactInfo : ContactInfo } “A contact must have an email or a postal address”
98. Making illegal states unrepresentable type Contact = { Name: Name Email: EmailContactInfo Address: PostalContactInfo } type Contact = { Name: Name ContactInfo : ContactInfo } type ContactInfo = | EmailOnly of EmailContactInfo | AddrOnly of PostalContactInfo | EmailAndAddr of EmailContactInfo * PostalContactInfo AFTER: Email and address merged into one type “A contact must have an email or a postal address” BEFORE: Email and address separate
100. Making illegal states unrepresentable type Contact = { Name: Name PrimaryContactInfo: ContactInfo SecondaryContactInfo: ContactInfo option } “A contact must have an email or a postal address” “A contact must have at least one way of being contacted” type ContactInfo = | Email of EmailContactInfo | Addr of PostalContactInfo
101. States and Transitions Modelling a common scenario
102. States and transitions State A State B State C Transition from A to B States and transitions Transition from B to A Transition from B to C
103. States and transitions Unverified EmailAddress Verified EmailAddress Verified States and transitions for email address Rule: "You can't send a verification message to a verified email" Rule: "You can't send a password reset message to a unverified email "
104. States and transitions Undelivered Out for delivery Delivered Put on truck Address not found Signed for States and transitions for shipments Rule: "You can't put a package on a truck if it is already out for delivery" Rule: "You can't sign for a package that is already delivered"
105. States and transitions Empty Cart Active Cart Paid Cart Add Item Remove Item Pay Add Item Remove Item States and transitions for shopping cart Rule: "You can't remove an item from an empty cart" Rule: "You can't change a paid cart" Rule: "You can't pay for a cart twice"
106. States and transitions Empty Cart Active Cart Paid Cart Add Item Remove Item Pay Add Item Remove Item States and transitions for shopping cart type ActiveCartData = { UnpaidItems: Item list } type PaidCartData = { PaidItems: Item list; Payment: Payment } no data needed
107. Modelling the shopping cart example type ActiveCartData = { UnpaidItems: Item list } type PaidCartData = { PaidItems: Item list; Payment: Payment} type ShoppingCart = | EmptyCart // no data | ActiveCart of ActiveCartData | PaidCart of PaidCartData Empty Cart Active Cart Paid Cart Add Item Remove Item Pay Add Item Remove Item
108. Shopping cart example initCart : Item –› ShoppingCart addToActive: (ActiveCartData * Item) –› ShoppingCart removeFromActive: (ActiveCartData * Item) –› ShoppingCart pay: (ActiveCartData * Payment) –› ShoppingCart Shopping Cart API Empty Cart Active Cart Paid Cart Add Item Remove Item Pay Add Item Remove Item
109. Shopping cart example let initCart item = { UnpaidItems=[item] } let addToActive (cart:ActiveCart) item = { cart with UnpaidItems = item :: cart.existingItems } Server code to add an item
110. Shopping cart example Client code to add an item using the API let addItem cart item = match cart with | EmptyCart –› initCart item | ActiveCart activeData –› addToActive(activeData,item) | PaidCart paidData –› ???
111. Shopping cart example let removeFromActive (cart:ActiveCart) item = let remainingItems = removeFromList cart.existingItems item match remainingItems with | [ ] –› EmptyCart | _ –› {cart with UnpaidItems = remainingItems} Server code to remove an item
112. Shopping cart example Client code to remove an item using the API let removeItem cart item = match cart with | EmptyCart –› ??? | ActiveCart activeData –› removeFromActive(activeData,item) | PaidCart paidData –› ???
113. Why design with state transitions? • Each state can have different allowable data. • All states are explicitly documented. • All transitions are explicitly documented. • It is a design tool that forces you to think about every possibility that could occur. Undelivered Out for delivery Delivered Put on truck Address not found Signed for
114. Review What I covered in this talk: • Ubiquitous language – Self-documenting designs • Algebraic types – products and sums • Designing with types – Options instead of null – Single case unions – Choices rather than inheritance – Making illegal states unrepresentable • States and transitions
115. Stuff I haven’t had time to cover: • Services • CQRS • The functional approach to use cases • Domain events • Error handling • And much more...
116. F# is low risk F# is the safe choice for functional-first development credit: @7sharp9 @MattDrivenDev Enterprise development F# on over 2 billion devices Mobile development Need to persuade your manager? -> FPbridge.co.uk/why-fsharp.html
117. Domain Driven Design with the F# type system DDD in F# resources fsharpforfunandprofit.com/ddd gorodinski.com tomasp.net/blog/type-first-development.aspx/ #fsharp on Twitter Contact me @ScottWlaschin FPbridge.co.uk