Re: [Boost-users] PropertyGraphConcept concept checking errors for an example implicit graph

Subject: Re: [Boost-users] PropertyGraphConcept concept checking errors for an example implicit graph
From: Jeremiah Willcock (jewillco_at_[hidden])
Date: 2010-06-26 11:30:35

On Fri, 25 Jun 2010, W.P. McNeill wrote:

> I'll look into implementing all the non-mutable graph concepts. Â That shouldn't be difficult. Â Are the following concept are consistent with each other?
>  *  Graph
>  *  IncidenceGraph
>  *  BidirectionalGraph
>  *  AdjacencyGraph
>  *  VertexListGraph
>  *  EdgeListGraph
>  *  PropertyGraph
> What should I specify for the traversal_category if I model all of them with a single object?

I believe they are all consistent.  Your traversal category should inherit 
virtually from all of the tags for the concepts that your graph provides. 
See the part in <boost/graph/grid_graph.hpp> that first mentions "virtual" 
for an example.

> I think also having a vertex property map is a good idea. Â Maybe I'll 
> make that a mutable color map, so there's an example of a writable 
> property map. Â Should I use associative_property_map as a base class? 
> Â (That's what's done in "The Boost Graph Library" section 15.3.2.)

I would use iterator_property_map (on a vector) with the vertex_index map 
that you have/are going to put in.

> I'll add Dijkstra's algorithm to the main() function.

OK.

> Have you looked at the source since commit f500721913f6728fc85d 
> "Complete Edge Weight Map Parameterization"? Â With that checkin I tried 
> to render all type definitions in terms of graph_traits<> and 
> property_traits<>.

Part of me says that it might be better to put your code in a namespace 
and then use the actual (internal) names of edge_descriptor, etc. to make 
the code shorter.  You can also, once you have a namespace, just make 
typedefs for "edge_descriptor" and such directly within the namespace so 
you can refer to them unqualified.  It's up to you whether you want to 
change that.

> I'll look at iterator_facade.
> 
> I'll combine implicit.cpp and implicit.hpp into just implicit.hpp. Â All 
> the valid expression functions should be inline anyway. Â I think having 
> a separate main.cpp makes things more readable.

It looks like you have done these in the latest checkin.

I would recommend having your source(), target(), etc. functions take the 
graph by const reference; although it doesn't matter to your code since 
your graph type is small, I think it would be a better model for other 
users (with larger graph types) that you not copy graphs around 
unnecessarily.

-- Jeremiah Willcock