It’s purely stylistic, there is no functional difference.
Consider the following:
type Foo = Bar
| Baz
| Bap
compared to this:
type Foo =
| Bar
| Baz
| Bap
The second example is a lot cleaner, and it’s immediately clear that the three things on the right side of the |
s are the constituents of the union.
Clearly, you wouldn’t add a leading |
when defining everything on one line:
type T = A | B