[boost] [BGL] Why not bf and df iterators (graphs vs trees)?

From: nicola (vitacolo_at_[hidden])
Date: 2006-06-23 14:53:50

Hi all,
the following is just a curiosity of mine, about which I would like to 
get some comment.

On browsing a few discussions about generic tree libraries, I got the 
impression that many people judge that pre-/in-/post-order iterators, 
and alike, are essential in any tree class or container (concept). My 
opinion is that a generic tree class (I refrain from calling it a 
"container", because in my mind a tree does not necessarily model a 
Container concept) should be designed much more after the BGL than after 
the STL containers.

One design choice in the BGL is that there are "visitor" functions, e.g. 
breadth_first_search() and depth_first_search(), but there are no 
"bfs_iterator" or "dfs_iterator". What are the motivations that lead to 
such choice? Might they be applied to trees as well?

As a simple example in favor of the former approach, a "BGL-like" tree 
data structure would allow a user to use a std::vector as a heap (by 
overloading the suitable functions).

I know that for the SoC several tree libraries have been proposed. Is 
there no room or interest for other proposals any longer? Is it possible 
to take a look at those proposals?

Regards
Nicola