[m-dev.] Alternative syntax for exporting things

[Fergus Henderson]

Peter Schachte wrote:
> 
> I've never been terribly happy with the :- interface and :- implementation
> sections of Mercury programs, as I'd rather write the pred/mode declarations
> for public predicates as I do the private ones:  right before the
> corresponding code.

The trouble with this is that it becomes difficult to separate the
module interface from the implementation.  So I would oppose any
such change, at least unless/until the proposal is accompanied
by a tool for extracting a human-readable versionof the module interface,
including comments.

> I know Fergus has a document somewhere specifying an
> alternative syntax for exporting things, but I find it a little bit too
> verbose.

"somwhere" = compiler/notes/module_system.html.

> So here's a simple alternative proposal.  One could optionally put `public'
> between the `:-' and the `pred', `mode', `inst', `type', etc in a
> declaration to specify that it is public, despite not appearing in an
> interface section. For example:
> 
> 	:- public pred append(list(T), list(T), list(T)).
> 
> 
> To allow abstract types, one could write something like
> 
> 	:- abstract type blah --> ....
> 
> or maybe
> 
> 	:- public abstract type blah --> ....

How about the following alternative?

	:- public type blah.
	:- type blah ---> ...

> This could also be extended to allow things to be reexported (ie, imported
> from one module that turns around and exports it) by just having the public
> pred/type/whatever declaration for an imported thing, with no actual
> definition in that module. 

Ugh.  I don't think that suggestion would work well.
(For example, what if you have to module-qualify the thing you
are re-exporting?)

I would much prefer a separate keyword for that.

> That's it.  This has the virtue of being quite simple, concise, and regular,

Agreed.

> and is still compatible with the current "section"-based approach.

You lost me there.
It seems quite different to the current section-based approach.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>   WWW: <http://www.cs.mu.oz.au/~fjh>  
Note: due to some buggy software and a (probably accidental)
denial-of-service attack, any mail sent to me between
	Tue Nov 25 20:00:00 UTC (6am Wed, local time)
and	Wed Nov 26 06:00:00 UTC (4pm, local time)
may have gone into the bit-bucket.  Please re-send it.