Re: modified-- strongly connected components(SCC)?

Amit Gupta wrote: 
I think, I should've been clear on this one (thanks for bringing it up)

1. If a single SCC contain 2M vertices of specific type, nothing is
reported for that component.

The other way to look at the desired output is: The result should be
similar to, if I instead do the following steps:
1. Run SCC without caring about any virtex-type.
2. From all the components I get from Step-1, prune those components,
which have vertices of type "that" more than M times.

I am surely missing the point, but I don't see the problem with filtering out the components that violates the constraint after the list of SCCs has been generated. As far as I understand, filtering out the non-valid components does not increase the complexity of the algorithm. Are you aiming at doing better than O(V+E)?

Luis
.