That is no doubt because _E is counted as a lowercase identifier by every other Haskell compiler. The use of _E is historical, that was what it was called in nhc98, and I never got round to changing the name. If someone wants to go through the libraries and rename it that would be fine :) Cheers Tom Neil Mitchell wrote:

Hi

...

One observation about _E: Haddock does not like it.

And having just checked, neither does GHC. Perhaps we shouldn't be using _E but something else - the definition:

data Box a = Box a

or

data Tup1 a = Tup1 a

both look good, and have some slightly obvious meaning. Tom - any reason not to prefer one of these?

Thanks

Neil

Malcolm Wallace wrote:

Whether this is actually a valid concern, given that the symbol can be hidden using the module system instead, is a different policy decision.

If I remember correctly, _E is defined inside the YHC.Primitive module. So unless someone deliberately imports that module (and they have no reason to do so) they will never know of the existence of the data type. Cheers Tom