Re: (infinite) sequences using recurrence relations

From: Eric Ford (eford_at_[hidden])
Date: 2001-08-07 07:30:32

> > So, yes, I beleive the container for a sequence generated via a
> > recurrence relation will need a new concept.  
> 
> 	No. Most people don't understand STL. Containers
> are NOT PART OF STL. They are provided merely as a way of
> creating some useful iterators. There are rules in the Standard,
> because the containers have to have requirements.

Ok, I'm definitely confused.  I guess I don't understand how to
reconsile "Containers are not part of stl" with "There are rules in
the standard [for containers?] because the containers have to have
requirements.".  It seems like the latter implies they are part of
STL.  Or maybe there's some precise definition to "are part of STL"
that I'm not getting.

The SGI STL web pages seem to imply that a list is a reversible
container which is a forward containe which is a forward container. 
Are they introducing additional concepts that are not in the standard
without mentioning that?