Re: "Multiple conformance"?

After encouragement to look at the problem again, I believe I found a
much more elegant solution in the form and I'm putting it out for
anyone else stumbling over this problem:

class a [G]

create
  make

feature
  make(input: G) is
    do
      feature1(input)
      feature2(input)
    end
  feature1(source: BOUNDED) is
    do
      ...
    end

  featue2(source: INDEXABLE) is
    do
      ...
    end

feature1 and will be able to handle input by its BOUNDED interface and
feature2 will handle it by its INDEXABLE interface.  It seems to me
this solution could be used in place of generic conformance restraints
since the compiler will determine at compile time whether input
conforms to BOUNDED and INDEXABLE when determining the feature
parameter types.

I always felt the '->' operator was ugly, I may stop using it unless I
find a reason that I'm missing.