Re: RTS graph and "temporal formulas"

["Dan'l Miller" <optikos@verizon.net>]

On Wednesday, August 21, 2013 7:19:19 PM UTC-5, Emanuel Berg wrote:
> I wrote:
>
> Only one more thing, for A, the "Moore-nomenclature response" is x
> <= 10, but transitions from A are at x < 5 and x > 11. Does this
> mean that the x < 5 transition is never made,

  You cross the line of answering the test question for you when you ask whether a transition will be made or not.

> because until x = 11, there is computation at A?

  Upon traversal of a transition that ingress into A (and only upon such traversal), there exists computation that consumes at most 10 clock ticks.  It is ambiguous in your professor's sloppy nomenclature whether that State-A computation is evoked upon start-up of the 7a FSA, such as we discussed above by traversal of an epsilon transition that initializes x to zero.  As stated, I would assume that the initialization of x is magic and that the State-A computation is not evoked upon start-up of the 7a FSA.

> Or does the A condition *hold* when
> it is OK to make transitions?

  Your terminology of "hold" is ambiguous.  If your "hold" means logically hold, then yes, the purple characterizations are absolute truth and always hold without exception:  upon every entry into a state in the 7a FSA that has a purple execution-time characterization, that characterization is never false.  I.e., the software always runs exactly as characterized, when so triggered.

  If your "hold" means pause, then no.  The purple characterizations of execution-time are ***not*** criteria that govern whether the entry-into-state computation executes or not, nor that govern when the entry-into-state computation executes.  Upon entry, that state's computation commences immediately and always executes to completion (as characterized).

> Or doesn't the A condition

  As I have said before, your confusion is because you are sloppy with your terminology.  The purple logical relations are not _a priori_ conditions (i.e., not criteria, not constraints governing a decision).  The purple logical relations are _a posteriori_ reports of how long the entry-into-state computation took to complete.

> impact flow,

  The purple characterizations do not directly affect a decision-making capability within the machinery that operates the FSA.  The purple characterizations may cause factual cascading ramifications (e.g., the passage of time as measured by clock ticks expended by the entry-into-state computation), where those cascading ramifications that are picked us by a sensor (e.g., reading the clock), where those sensor readings in turn affect whether a transition is traversed, as governed by the green stimulus-constraints on (some) transitions.  Impact flow?  Directly, no.  Indirectly, perhaps, and analysis thereof is what your professor is requesting in the exam questions 7.1 and 7.2 and subbullets thereof.

> it just tells something about the system at state A?

  Yes, as I have been saying, the green logical relations are _a posteriori_ reports of how the entry-into-state computation turned out.

> Or does computation "fork" so there are numerous processes all around
> the graph?

  No.  FSAs are DFAs (deterministic nonforkingly, despite the possibility of single epsilon transitions borrowed from NFAs, as we discussed above).  FSAs are not a GLR parser, which does in fact have the exploratoration-of-hypotheticals forking behavior that you (incorrectly for FSAs) devise.