Boost logo

Boost :

From: jsiek_at_[hidden]
Date: 2000-01-07 05:27:55


I've just uploaded an update of the experimental graph code. It
checks out on VC++ 6.0sp3 and g++ 2.96.

Dave Abrahams writes:
> > I've been doing some experimenting with some graph classes... most
> > of the work is in figuring out what will work with VC++ :(
> > I'll finish up those experiments tonight and post them.
>
> Great!
>
> > Dietmar, have you had a chance to do some graph implementation
> > work?
>
> Anything I can do?

Since you're interested in the reversable graph, you might try coding
that up. Perhaps if you model it after the adj_list class, but with
a vector of edge-lists for the in-edges and a vector of edge-lists
for the out-edges.

> > > Also, something I've found to be very useful: a reverse graph adapter. If
> > > you have a directed graph where the incoming edges are available for a
> node,
> > > you can wrap it to produce an identical graph with all of the edge
> > > directions reversed. I hope this will be possible with our library.
> >
> > Hmm... in_edges() isn't in the current interface. I'll try to make a
> > place for it. Most algorithms don't need it, so it will probably be a
> > singleton requirement in a new conept... any ideas on a name for
> > this kind of graph?
>
> what about:
> ReversibleGraph
>
> Is that the right spelling? It looks right to me. I tried "reverseable",
> too.

Yes, that is the right spelling.

>
> On the other hand "BidirectionalGraph" might be more in keeping with the
> standard.

Nod.

Cheers,

Jeremy


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk