Declarative programming states what result or transformation is wanted while leaving low-level execution details to the compiler or supporting library. Imperative code instead specifies how: create intermediate data, iterate, mutate, and manage control flow explicitly. The declarative style is often shorter, clearer, and easier to compose. section 2.1.1 section 7.2.1
IEnumerable<int> GetEvenNumbers(IEnumerable<int> numbers) #1
{ #1
var output = new List<int>(); #1
foreach (var number in numbers) { #1
if (number % 2 == 0) #1
output.Add(number); #1
} #1
return output; #1
} #1
let getEvenNumbers = #2
Seq.filter (fun number -> number % 2 = 0) #2
Compare a manually built even-number list with the compact F# sequence filter. The declarative version names the desired operation without exposing the loop or accumulator. section 2.1.1
A useful way to design a declarative solution is to describe the data transformations first. For a results query, filter away wins, count them by away team, then select the largest count. A pipeline expresses those stages directly: each operation produces a new result set for the next operation. section 7.2.1
Declarative pipelines make several practical properties visible:
- Immutable data avoids repeated runs accumulating changes in a shared collection. section 7.2.1
- Expressions expose meaningful transformations more clearly than statements returning unit. section 7.2.1
- Independent stages can be reordered or extended with less risk of disturbing other stages. section 7.2.1
- Libraries provide declarative models in otherwise imperative languages, including LINQ in C# and Streams in Java. section 2.1.1
Modeling a view declaratively
Declarative interfaces describe structure rather than manually constructing every UI operation. XAML represents views through markup, while F# can model an entire view in code and still support conditionals and loops directly. ASP.NET Razor occupies a hybrid position by allowing C# inside markup. section 13.4.1
