(j3.2006) derived types with type parameters are different
Tue Jul 29 16:46:46 EDT 2008
>1) Too drastic: Disallow parameters on SEQUENCE types
But this is equivalent to
Disallow SEQUENCE on parameterized types
which sounds reasonable to me.
If it causes significant loss of functionality, what exactly is the
functionality that we lose?
1) Well, there's the ability to specify the same type in different places.
But that's what this thread has been about, and we don't seem to have settled on
exactly how one does that.
2) There's also the ability to specify a storage sequence for elements of that
type. But so what? According to 188.8.131.52 a nonpointer scalar object of
a type with type parameters occupies a single unspecified storage unit
whether or not it is a sequence type, so this is no functionality at all.
Specifically, for derived types,
case (5) on page 416 covers nonpointer scalar objects
of sequence type with no type parameters and case (6) picks up the leftover
cases, namely sequence or nonsequence types with type parameters.
More information about the J3