> -----Original Message-----
> From: gotoken / hanare00.math.sci.hokudai.ac.jp
> [mailto:gotoken / hanare00.math.sci.hokudai.ac.jp]On Behalf Of GOTO
> Kentaro
> Sent: Monday, January 08, 2001 07:52 AM
> To: ruby-talk ML
> Subject: [ruby-talk:8878] good container?
> I see what you say just now.  Thank you Christoph.  Well, such
> development can be done efficiently by release-test-improve loop.  So,
> I'd like to wait RAA-entry(s) of containers written by anyone.  If
> cachy one appears for everyone, Matz adopt that as a standard library.
This seems like a pragmatic and probably best course of action.
> >class ContainerClass
> >ZERO  =  0 # or something similar like the hypothetical Numeric::ZERO 
> >def   def sum(if_empty = type::ZERO)
> >..
> >end
> >...
> >end 
> >
> >This way you can override the generic ZERO in a specialized
> >ContainerClass 
> 
> I think operations like sum is a kind of cross-hierarchical spec of
> container and it should be provided as mix-in, perhaps.
Agreed ..
> Cle and me had a series of discussions about container hierarchy in
> the context of how to realize sutably reverse ordered version of each.
> You can find them in some threads, e.g.:
> 
> `New feature for Ruby'
> http://blade.nagaokaut.ac.jp/cgi-bin/vframe.rb/ruby/ruby-talk/407?404-621
Interesting thread - funny this is (almost) exactly what people are 
discussing in the current ``Backward Range implementation'' thread.

Thanks!
Christoph