`On Thu, 3 May 2001, Eric S. Raymond wrote:> Alexander Viro <viro@math.psu.edu>:> > I'm not talking about connectedness of the thing. However, I suspect that> > graph has a small subset such that removing it makes it fall apart.> > Um.  So how does that help?First of all, it reduces the complexity of finding the best validapproximation (mind you, I'm not saying that it's the problem youneed to solve, just that you don't need anything close to 3^1976even for _that_).And simple variation of the full thing can be used for "find avalid approximation within given distance".Full (and dumb) variant first:	for all combinations of values of core variables		for each peripherial group			find the best valid approximation within that group,			given the values of core variables.The cost is 3^(size of core) * 3^(size of maximal peripherial group) *number of peripherial groups * size of maximal periph. group * maximalnumber of constraints for periph. group.Again, I'm _not_ suggesting to do that. However, limited variant of theproblem (find valid approximation within 10 changes from given or complain)is much easier.Eric, could you point me to a place where I could grab the currentset of constraints? In any case I have a very strong gut feeling thatseparation the variables into core and peripherial is essential partof data and ignoring it makes problem much harder than it really is.I'd like to see how large the core actually is and how large periph.groups are.-To unsubscribe from this list: send the line "unsubscribe linux-kernel" inthe body of a message to majordomo@vger.kernel.orgMore majordomo info at  http://vger.kernel.org/majordomo-info.htmlPlease read the FAQ at  http://www.tux.org/lkml/`