Difference between a Seq and a List in Scala

In Java terms, Scala's Seq would be Java's List, and Scala's List would be Java's LinkedList.

Note that Seq is a trait, which is equivalent to Java's interface, but with the equivalent of up-and-coming defender methods. Scala's List is an abstract class that is extended by Nil and ::, which are the concrete implementations of List.

So, where Java's List is an interface, Scala's List is an implementation.

Beyond that, Scala's List is immutable, which is not the case of LinkedList. In fact, Java has no equivalent to immutable collections (the read only thing only guarantees the new object cannot be changed, but you still can change the old one, and, therefore, the "read only" one).

Scala's List is highly optimized by compiler and libraries, and it's a fundamental data type in functional programming. However, it has limitations and it's inadequate for parallel programming. These days, Vector is a better choice than List, but habit is hard to break.

Seq is a good generalization for sequences, so if you program to interfaces, you should use that. Note that there are actually three of them: collection.Seq, collection.mutable.Seq and collection.immutable.Seq, and it is the latter one that is the "default" imported into scope.

There's also GenSeq and ParSeq. The latter methods run in parallel where possible, while the former is parent to both Seq and ParSeq, being a suitable generalization for when parallelism of a code doesn't matter. They are both relatively newly introduced, so people doesn't use them much yet.

A Seq is an Iterable that has a defined order of elements. Sequences provide a method apply() for indexing, ranging from 0 up to the length of the sequence. Seq has many subclasses including Queue, Range, List, Stack, and LinkedList.

A List is a Seq that is implemented as an immutable linked list. It's best used in cases with last-in first-out (LIFO) access patterns.

Here is the complete collection class hierarchy from the Scala FAQ:

https://i.stack.imgur.com/2fjoA.png

Seq is a trait that List implements.

If you define your container as Seq, you can use any container that implements Seq trait.

scala> def sumUp(s: Seq[Int]): Int = { s.sum }
sumUp: (s: Seq[Int])Int

scala> sumUp(List(1,2,3))
res41: Int = 6

scala> sumUp(Vector(1,2,3))
res42: Int = 6

scala> sumUp(Seq(1,2,3))
res44: Int = 6

Note that

scala> val a = Seq(1,2,3)
a: Seq[Int] = List(1, 2, 3)

Is just a short hand for:

scala> val a: Seq[Int] = List(1,2,3)
a: Seq[Int] = List(1, 2, 3)

if the container type is not specified, the underlying data structure defaults to List.

In Scala, a List inherits from Seq, but implements Product; here is the proper definition of List :

sealed abstract class List[+A] extends AbstractSeq[A] with Product with ...

[Note: the actual definition is a tad bit more complex, in order to fit in with and make use of Scala's very powerful collection framework.]

As @daniel-c-sobral said, List extends the trait Seq and is an abstract class implemented by scala.collection.immutable.$colon$colon (or :: for short), but technicalities aside, mind that most of lists and seqs we use are initialized in the form of Seq(1, 2, 3) or List(1, 2, 3) which both return scala.collection.immutable.$colon$colon, hence one can write:

var x: scala.collection.immutable.$colon$colon[Int] = null
x = Seq(1, 2, 3).asInstanceOf[scala.collection.immutable.$colon$colon[Int]]
x = List(1, 2, 3).asInstanceOf[scala.collection.immutable.$colon$colon[Int]]

As a result, I'd argue than the only thing that matters are the methods you want to expose, for instance to prepend you can use :: from List that I find redundant with +: from Seq and I personally stick to Seq by default.