Re: [Boost-users] Re: Iterators and maps [was Re: Where should STL suggestionsgo?]

From: Jim Lear (jim.lear_at_[hidden])
Date: 2004-12-14 11:29:54

David Abrahams wrote:
> Jim Lear wrote:
> > So in the absolute most abhorently poor code example (let's just call it 
> > "meta-code"), the map operator would behave like:
> >
> >     data_type &operator[](const iterator &i) { return i->second; }
> >
> > Am I just too ignorant to make any sense? :-)  Maybe I'll crawl back into my 
> > hole. :-)
> >   
> You're making sense, but consider this: why should the map get involved
> in that operation at all when you can do it all with the iterator?
> Seems like a design flaw to get another object involved needlessly.

Not only might it be a design flaw, but it may invite inefficiency.  But
really it's just consistency with other languages which iterate over the
keys (like awk) and code clarity.  Someone unfamiliar with STL would find
the following obvious:
     for (m_type::iterator i=m.begin();...)
        for (mi_type::iterator j=m[i].begin();...)
             cout << m[i][j] << endl;
 But they may find
     for (m_type::iterator i=m.begin();...)
         for (mi_type::iterator j=i->second.begin();...)
              cout << j->second << endl;
obtuse.  Perhaps folks intimately familiar with iterators may find this is
hard to believe.  :-) An alternative question might be, why should one who
uses maps and vectors also be required to be familiar with the intricacies
of iterators?  To me, that seems like a design flaw.  I have never encountered
another language that requires the user to understand iterators to the extent
STL does.  Ease of use and learning ought to be a priority in the design
of these objects, to the extent it is efficient and safe.  To use iterators
in lieu of keys or indices seems to me like a great simplification for users.
 It completely obviates the need to understand iterators in this case (while
still allowing existing iterator functionality).
-- 
Jim Lear