Skip to main content
5 events
when toggle format what by license comment
Apr 14, 2015 at 11:53 comment added sdenham @ArtB : You should modify the question to include the update to the problem statement you have given here. Also, it may help to state that these are directed edges (if that is the case), and whether a cycle would be considered an error in the graph, rather than just a situation the algorithm needs to handle.
Apr 13, 2015 at 21:07 comment added sdenham @ArtB : Thank you for your clarification of the problem. As you have said there is no more than one edge between any two vertices, and are apparently ruling out paths with cycles (or at least more than one pass around any cycle), then at least we know there are not literally an infinite number of possible valid combinations, which is progress. Note that knowing how to enumerate all the possibilities is not the same as saying you must do it, as it could be a starting point for an making an argument for correctness, which in turn can guide testing. I will give it more thought...
Apr 13, 2015 at 19:12 comment added Sled @sdenham How are you going to enumerate something that literrally has a infinite number of possible valid combinations? I was hoping to find something along the lines of "these are the trickiest graph structures that will often catch bugs in your implementation". I understand the domain well enough as it is simple: For all vertices N in forest F, for all vertices M, in F, such that if there are any walks between N and M they all must either use only edges labelled 'conflict' or 'requires'. The domain is not the issue.
Apr 13, 2015 at 18:38 comment added sdenham The questioner has written that they are unable to tell whether they have considered all cases, which implies that don't have a way to enumerate them. Until they understand the problem domain well enough to do that, how to test is a moot question.
Apr 13, 2015 at 18:06 history answered Macke CC BY-SA 3.0