Why not create two different Iterator types: Standard_Iterator and
Reversed_Iterator (Or Cat_Iterator and Dog_Iterator)?
Each would have its own First, Last, Next, and Previous subroutines.
Perhaps a Flip function would be useful too.

-CRM

"Mark Lundquist" <mlundquist2@attbi.com> wrote in message
news:dcKQ7.23845$Yy.297532@rwcrnsc53...
>
> "Jeffrey Carter" <jeffrey.carter@boeing.com> wrote in message
> news:3C114702.98662A90@boeing.com...
> >
> > If we can't agree on the basic properties of lists we'll never get
> > anywhere. Perhaps we need an appeal to authority here.
>
> Look, you guys both know what a doubly-linked list is; you don't need some
> double-dome to weigh in with a ruling on it!
>
> You are having an argument about nomenclature, not about the basic
> properties of anything.  You just think you are because you have befuddled
> each other with bad arguments.
>
> Jeffrey -- of course "sequence" (extrinsic ordering) is a fundamental
> property of any kind of a linked list.  I don't think anyone is confusing
> the data structure in question with a set, bag, heap, or any other kind of
> intrinsically ordered thing.  Your point is irrelevant to the nomenclature
> question.  Ted doesn't deny that a list has direction, he's saying that a
> doubly-linked list has two directions, and for some reason he feels
strongly
> about any kind of preferential scheme that would seem to establish one end
> or direction as secondary or relative (like "Normal" vs. "Bass_Ackwards"
> :-).
>
> But Ted, what's the big whoopie deal about this, anyway?  Who cares if the
> names have a "directional bias", as long as the semantics are clear?  The
> important thing is the relationship between the names you choose for the
> extremities and the names you choose for "direction", right?  So if the
> extremities are "Bow" and "Stern", then the directions had better be
> "Forward" and "Aft".  That's why "Head/Tail" is kinda bad -- with a na�ve
> choice for the direction names, like "Forward/Reverse", even the
originator
> of the naming scheme probably wouldn't be able to keep them straight :-).
> But you have to start somewhere, and everybody knows it's arbitrary which
> ends you call "First" and "Last".  I don't buy the argument that a
> preferential naming scheme entails a loss of flexibility or that it
obscures
> the property of bidirectionality.
>
> Cheers,
> -- mark
>
>
>
>