Re: [stack] Self Types

I can¹t extract the paragraph from your link, so I have to paraphrase:
You mention the difficulty of typing what in Cat is dup apply, or in Joy
dup i. I have occasionally defined a combinator x == dup i, and I chose
that name because it resembles the y combinator. In particular, x can
do whatever y can do: implement recursive execution without a recursive
definition. See the section ³Self-application² in my note:

http://www.latrobe.edu.au/philosophy/phimvt/joy/jp-nestrec.html

My recollection is that there are some difficulties about typing the
y combinator in lambda calculus languages, but don¹t ask me for details.
Presumably the same holds for the y combinator in concatenative
languages, and by extension, for the x combinator in Joy and hence
for dup apply in Cat. Hope this helps.

  - Manfred



On 9/6/07 4:24 AM, "Christopher Diggins" <[email protected]> wrote:
> 
> My previous plan for dealing with recursive (self-referential) types
> was to use labels. I am currently looking at using "self" types, which
> are more limited, but I think sufficient.
> 
> http://cdiggins.com/2007/06/07/self-types/
> 
> Christopher
>  
>     




[Non-text portions of this message have been removed]