ex4: Should our indexing strategy work with all monitors

[LordHoto]

Hallo,

the exercise sheet suggests that our implementation should be compatible to *all* monitoring schemes. Now the template you gave us takes a FSM as monitor template. Due to that we implemented our indexing strategy with respect to FSM properties now and thus we cannot plug our code into the AllTests class for example. That makes us not be as generic as the builtin strategies StrategyB and StrategyC.

Now the question is: Should we generalize the symbol based indexing scheme to work also with monitors which are no FSM?

// Johannes

[ericbodden]

Hello Johannes.

Now the question is: Should we generalize the symbol based indexing scheme to work also with monitors which are no FSM?

No, that is not required. The strategy should minimally work for the two properties provided. Ideally it should also work for FSMs in general. Anything beyond that is not required for this exercise (although it would still be interesting to think about how that could be achieved).

[LordHoto]

No, that is not required. The strategy should minimally work for the two properties provided. Ideally it should also work for FSMs in general. Anything beyond that is not required for this exercise (although it would still be interesting to think about how that could be achieved).

Ok I was a bit confused since on the exercise sheet it states:

It should be correct for any MOPBox-based runtime monitor.

which makes me think that we should have our symbol based indexing work with a generic IMonitorTemplate and IMonitor. With these interfaces the problem however is that we cannot obtain any information on "OutTrans", since IMonitor does not supply the list of the currently accepted symbols. I do not see any easy solution for that though .

[ericbodden]

which makes me think that we should have our symbol based indexing work with a generic IMonitorTemplate and IMonitor. With these interfaces the problem however is that we cannot obtain any information on "OutTrans", since IMonitor does not supply the list of the currently accepted symbols. I do not see any easy solution for that though .

Yes, in this case one would have to introduce another abstraction again (similar to the (co)enabled-sets that JavaMOP fetches from the logic plug-ins). That would be too much to ask for the purpose of this exercise.

But you are right about the sentence in the exercise sheet: this could have been stated more clearly. Sorry about that.