Each primitive module should export an eponymous type

[crusso]

To avoid cylic definitions, requires a cycle breaking Types.mo (useful anyway) or motoko's special mo:prim to export all primitive types under their own name.

[crusso]

Actually, I'm now wondering if it would in fact be better to just uniformly call these Type, (Word32.Type etc.) to avoid the circularity, redundant Word32.Word32 and in anticipation of functors. @rossberg @nomeata @matthewhammer

[nomeata]

motoko's special mo:prim to export all primitive types under their own name

That seems reasonable and non-intrusive

[rossberg]

Hm, I'm sceptical. IME, it is more cryptic and less natural to most folks, and it gets in the way of import {...} = "..." (which we should add). OTOH, functors are not going to be relevant for Motoko anytime soon. And: not even SML does it. ;)

I remember a discussion on TC39, where some folks argued that this pattern demonstrates why ML modules are a practical failure...

[nomeata]

What are you sceptical about? Exporting Int.Type, Int.Int or anything at all?

[rossberg]

@nomeata, I mean Int.Type.

[crusso]

Fine, I'll go for the prim extension and add eponymous type components. We can always add *.Type later if it ever becomes useful