A calculated property takes in two things: the name of the property and the expression that would produce the desired result. Technically, this is a hashtable with n (or Name) and e (or Expression) being the two name-value pairs. The Expression in itself is a script block. When broken down into lines, the query looks like this:

Get-ChildItem -File -Recurse | Select-Object Name, LastWriteTime, @{
    Name = "Size"
    Expression = {
        [math]::Round($PSItem.Length/1MB, 3)
    }
}

Since we separated the Name and the Expression by a line, we don't need the semicolon anymore.

Calculated properties come in handy in situation where you don't really want to create a new script, but instead, want some information at the terminal itself, as though running a query. The value of Name is the name of the property, and the values in the column are determined by the expression.