Hazard checking algorithm version 1.3 10/01/2002

1. [Done] Color all states with R or F based on the Boolean evaluation of the node under consideration.

2. [Done] Speed Independent edge stabilization. Find the state->state transitions where the output changes and see if the path from my node of interest to the output can propagate. If so, then color the edge with a 0 or 1, depending on the Boolean evaluation. Store these stabilized edges in a linked list for propagation later.

3. [Done] Find points where the state graph has state-to-state transitions from R->F or F->R (this is where there is a change in evaluation on my node under consideration). Store these in a separate linked list. For each of these, do the following:

4. [Done] Find the edge transition at which this occurs (i.e. a+, b-, etc.). Store this in the linked list.

5. [Done] Set this as the reference point.

6. [Done] Find the maximum delay from this reference point to the node of interest in my decomposition. If more than one path, take the outside bounds.

Notes: Sense ignored in decomposition i.e. if the reference transition is a+ then a timing must be calculated for any a+ or a- input found in my decomposition.

Assume the reference transition is causal to the unknown node changing evaluation. Then, the path from the reference input to the node of interest is always enabled and the time delay through my decomposition can be computed directly from reference transition to node.

7. [In progress] Find the zone associated with the state following (or preceding?) the reference transition and make a copy of it. For each edge leaving this state:

8. Find the causal event for this transition (for now, assume it’s the previous transition).

9. Add this event to the zone.

10. Modify zone with new <lower upper> information.

11. Canonicalize zone.

12. Extract new timing information from zone.

13. Compare the minimum elapsed time in the zone with the maximum delay through my decomposition.

14. If the minimum elapsed time is equal to or greater than delay through my decomposition, color this edge with a stable color (1 if color was Rising, 0 if color was Falling). Go to step 3, else go to step 7.

15. If I reach a termination point, I cannot stabilize the edge so leave it colored as is. Go to step 7. (A termination point is when I find an edge enabled in the opposite direction of my reference transition or I reach an edge previously stabilized.)

16. Propagate the stable edges from the SI hazard check and the timed hazard check.

17. Propagate the stable edges from the start state.

18. Check for and report coloring inconsistencies.