[bitc-dev] Current thoughts on mixfix

[wren ng thornton]

On 9/4/10 9:20 AM, Jonathan S. Shapiro wrote:
> On Fri, Sep 3, 2010 at 9:29 PM, wren ng thornton<wren at freegeek.org>  wrote:
>
>> In particular, it's fine to define
>> the *function* _+_ by saying it's equal to add (and then
>> specifying/relying on inlining to remove the indirection); but having
>> _+_ belong to a separate ontological category from add is just begging
>> for problems.
>
> I tend to agree. Unfortunately there is a moderately bad problem with this:
> it ties the name of the syntactic element to the name of the function.
>
> So now, if I want to define a completely new syntax in which _+_ means
> something else entirely, I get tied up in namespace issues...

That's part of the reason I brought up Dyna. Since you're already 
planning on having modules that can be "opened" into the local context, 
doesn't that subsume syntax tables? A syntax is just (the names for) a 
collection of operations; a module is just a collection of 
operations[1]. If you want _+_ to mean something different in the 
current context, then just define a local _+_ and don't open import the 
_+_ from other modules--- use the non-local versions with qualified 
names instead. If you want a local non-qualified name you can always 
define something like _<+>_ = M._+_ or _++_ = M._+_ or whatever.


[1] In both cases taking "operation" to mean functions, constants, types 
(which are just type-level functions or constants),...

-- 
Live well,
~wren