[Ur] last type decl wins?

Adam Chlipala adamc at impredicative.com
Sat Apr 16 11:21:39 EDT 2011


Vag Vagoff wrote:
>> No. I often use that style, and I even prefer it to the alternatives 
>> that you suggest. I like being able to read definitions in a simple 
>> linear order, so that it's always clear what each symbol refers to.
>
> How about to enable such redefitions in local defs but not on 
> toplevel? Will it be optimal solution?
> Large modules may be easily turned to incomprehensible entangled mess 
> due to several different equally named definitions.

I think of almost every Haskell program as an "incomprehensible 
entangled mess," because of implicit mutual recursion spread across 
files!  Being forced to choose different names for similar functions 
with different scopes leads to longer and harder-to-understand names.

I have no plans to change Ur/Web's treatment of shadowing, as I believe 
the way it's done now is already best for me.  I might still accept a 
patch that someone else comes up with, adding an optional class of warnings.



More information about the Ur mailing list