Should it be explained more thoroughly that a a function returns a projection?

Let's say we have a list of list and want to select the sublist where a string starts with a letter.

So, given the JSON
```json
[
[
"ab",
"cd"
],
[
"ef",
"gh"
]
]
```

One could do:
```none
[?contains([].starts_with(@, `a`), `true`)]
```

But, in order to have a consistent notation, I was trying to make the `contains` function a dot notation too, so:

```none
[?[].starts_with(@, `a`).contains(@, `true`)]
```

which does not work.

I figured out it as because `starts_with` returns a projection, so it can be achieved resetting it:

```none
[?[].starts_with(@, `a`) | contains(@, `true`)]
```

But I feel, for a new comer that does not know that well about projections and resetting them, it should be explained in the specification in more details, shouldn't it?