develooper Front page | perl.perl6.porters | Postings from January 2000

Re: New Year! New Topaz Snapshot! New Project Home!

Thread Previous
From:
William Setzer
Date:
January 3, 2000 14:52
Subject:
Re: New Year! New Topaz Snapshot! New Project Home!
Message ID:
lzyhfgusr8i.fsf@babylon5.unity.ncsu.edu
Chip Salzenberg <chip@valinux.com> writes:
: 
: 2. There's a new Topaz snapshot available: topaz-0.09.tar.gz.

I have a couple questions/suggestions:

*) If for any given parent class there is only one child class
   inheriting from it, is there an advantage to separating them?  The
   extra level of abstraction seems a bit confusing to me.  Example:

   Scalar <- FinalScalar <- PoolScalar

   Will there ever be another class directly inheriting from Scalar or
   FinalScalar?  If not, then why not fold the classes together and
   minimize the "where did this method come from/what's my inheritance
   path" confusion?

*) It might be a great help if the distribution's file hierarchy
   somehow mirrored the class hierarchy.  An incomplete hypothetical:

     > ls -R Value
     Aggr/     Code/     IO/       Scalar/   Value.cc  Value.hh

     Value/Aggr:
     Aggr.cc  Aggr.hh

     Value/Code:
     Code.cc  Code.hh

     Value/IO:
     IO.cc  IO.hh

     Value/Scalar:
     Final/         Scalar.cc      Scalar.hh      ScalarUtil.cc  ScalarUtil.hh

     Value/Scalar/Final:
     FinalScalar.cc  FinalScalar.hh  Pool/

   This leverages people's perl class thinkspace and graphically
   points out both strange object inheritance (e.g. Glob) and single
   parent/child inheritance.

Is is appropriate at this point to discuss inheritance paths, or
will things change around a bit?


William

Thread Previous


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About