Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#15 Incorrect operational semantics

open
Vadim Gurov
5
2006-07-21
2006-07-21
Vadim Gurov
No

Unfortunately there are some problems preventing to
make it as simple as possible.
following precondition:
-there is a state machine 1 withe a state sx and sy ,
and the transition
e1 from sx to sy:
start --> sx --e1-->sy
- there is a state machine 2 with a start node, the
states s1 and s2, and a transition e1 from s1 to s2:
start --> s1 --e1-->s2

scenario 1.:
- state machine 2 is defined as a separate state machine
- state machine 2 is included into state sy of state
machine 1
- state machine 1 is in state sx
- the event e1 is send to state machine 1
- state machine 1 enters sy
- state machine 2 is entered, automatically transits
from start to s1 and transits with e1 from s1 to s2
+finally the state machine 1 is in sy and machine 2 is
in s2

scenario 2.:
- state machine 2 is modeled directly in the sy state
of machine s 1
- state machine 1 is in state sx
- the event e1 is send to state machine 1
- state machine 1 enters sy
- state machine 2 is entered on start node and e1
transits it to s1
+so the final state is machine 1 is in sy and machine 2
is in s1

I would expect a similar behavior in both cases,
preferably the behavior
of scenario 1.

Discussion

  • Vadim Gurov
    Vadim Gurov
    2006-07-21

    Logged In: YES
    user_id=240847

    An other problem is the state "hopping":
    if a state A contains the states B and C
    B and C contain a state D each
    the state machine has entered state A-C-D somehow
    at the next event the machine is in state A-B-D unexpectedly
    (probably the other way around, maybe depending on which is
    defined first)
    do the states have to be unique in name?