Re: [boost] [Review] Boost Graphs Library (GGCL)

From: Jens Maurer (Jens.Maurer_at_[hidden])
Date: 2000-08-31 16:49:03

Part four.

Sorry, I'm tired, so no code inspection.

General
-------

 - A single size_type as shown in adjacency_list (doc 8.1.5) for both
edge and vertex numbers is not enough when edge and vertex containers
differ in their size_type.

 - It looks like container_gen<> is a complete replacement for 
template template parameters.  Adding further specializations
under "namespace std" rules is no problem. But I still think
that template template parameters are better (more localized
information).

Documentation
-------------

8.1.1: Please add that you can define your own selectors as well and
that the given list is not exhaustive.  8.1.2 does say that
I can have my own selectors. Please add a reference to 8.1.7.

8.1.2: The "Space Complexity" section is empty.

8.1.2, Time Complexity: The description should be more uniform
in their reference to the two different basic container variants.
edge() talks about Sequence and AssociativeContainer, remove_edge()
talks about sequence-based and set-based.  Is there any difference?
clear_vertex() has yet another saying: "associative container based".

8.1.2, vertex_iterator::operator++():  Is this always equivalent to
incrementing a pointer, even when VertexList is a std::list?
I don't think so.

8.2: std::pair has "first" and "second" members, not "first" and "last".

8.2: I am not sure what the ValueType and Diff template parameters
are there for.  The "template parameters" section doesn't say
what other reasonable things I could do with them, so why aren't
they just typedef's within edge_list<> ?

12.1.3: "hundredths of seconds" is a crazy unit.  Can't we have
milliseconds instead?

Jens Maurer