[sorry, I've lost the original post] On Thu, Feb 23, 2006 at 04:47:37PM -0800, John Meacham wrote:
On Thu, Feb 23, 2006 at 11:36:34AM +0100, Wolfgang Jeltsch wrote:
The remaining problem is that using "type" for every type is misleading, since "type" is otherwise used only for aliases. Maybe it would be better to use something like "typealias" for type aliases, since using "type" for declaring type aliases is misleading anyway.
I'd be in favour of H' supporting "typealias" as well as "type". This would lead the way, via compilers giving warning for uses of "type", to dropping support for "type" more painlessly in a future standard. Too much code has variables called "typ"... I think the same with "data" and "datatype" would be nice too. I also have a number of variables called "dat" dotted around. We could still use "type" in export lists, if desired, in the same way we use can still use things like "export" in FFI exports (i.e. you might have to export "variable type" rather than an unqualified "type"). Thanks Ian