Hi Robert.

Okay, I think I mostly understand it now, but I'm still not sure about one
thing:

When we are listing prime compatibles:

If there are two states that could be used, each of which covers a pair, and
one of the states in each pair is shared between them, do we need to list
the states that differ as well, or do we only do that when the state
covering a pair (or clique) implies another covering state?

e.g., if we could make states XY and XZ (no implications), do we list Y and
Z again on their own?

Thanks,

Ben

On Mon, Nov 24, 2008 at 6:02 PM, Robert Dick <dickrp@northwestern.edu>wrote:

> > Sorry to keep pestering you with questions, but here goes:
> >
> > I remember a lot about how to do the FSM minimization from the example in
> > class - but I am fuzzy on parts of it.
> >
> > Once we have listed the prime compatibles using the implication chart,
> and
> > are to draw the nodes and connections (with implications if there are
> any),
> > do we ever draw nodes more than once? (e.g., when state X and Y can be
> > covered by a single state but that implies YZ, should we draw X twice,
> once
> > connected to Y and once on its own?)
>
> Each node will appear only once, and each will correspond only to a single
> state in the original FSM. Then start listing compatibles, starting from
> the
> maximal compatibles and adding subsets when this is necessary in order to
> cover a state with reduced implications.
>
> > Or do we just have which cliques we circle show that?
>
> Right, and which grouping you enter into your binate covering table.
>
> > Also, if we can cover several pairs of states but it would give the same
> > amount of states whether certain states are paired or left separate,
> should
> > we go ahead and cover the pairs? (I was thinking... yes if it would make
> > the FSM's behavior more specified, otherwise it makes no difference?)
>
> In this particular case, no. By leaving more Xs in the final machines, you
> may have more opportunity for minimizing the resulting state variable and
> output functions.
>
> Best Regards,
>
> -Robert Dick-
>
>
Received on Mon Nov 24 18:29:28 2008