[ACT-R-users] Act-r time question

[Emond, Bruno]

There it is.
Thanks for the clarification as to where the time is spent.
Bruno

On Nov 4, 2004, at 13:53, Dan Bothell wrote:

>
>
>
> --On Thursday, November 04, 2004 12:43 PM -0500 "Emond, Bruno"
> <bruno.emond at nrc-cnrc.gc.ca> wrote:
>
> > I am a little bit puzzled by the following act-r trace.
> >
> > The productions are fairly simple.
> > ?Recall-Text-Location calls +retrieval>
> > ?Test-Recall-Not-Successful tests that
> >    =retrieval> isa error condition failure
> >    the it changes the goal to find the visual location
> >    of a text
> > None of these productions do pm calls.
> >
> > The productions ?Find-Text-Location and ?Attend-To-Text
> > call respectively +visual-location and +visual
> >
> > My question is what has appended to the time spent on
> > ?Recall-Text-Location
> > which shows when ?Test-Recall-Not-Successful fires?
> >
> > I am running act-r/pm 2.2
> >
> >  Time  0.000: Vision found LOC103
> >  Time  0.000: ?Recall-Text-Location Selected
> >  Time  0.050: ?Recall-Text-Location Fired
> >  Time  2.050: ?Test-Recall-Not-Successful Selected
> >  Time  0.100: ?Test-Recall-Not-Successful Fired
> >  Time  0.100: ?Find-Text-Location Selected
> >  Time  0.150: ?Find-Text-Location Fired
> >  Time  0.150: Module :VISION running command FIND-LOCATION
> >  Time  0.150: Vision found LOC104
> >  Time  0.150: ?Attend-To-Text Selected
> >  Time  0.200: ?Attend-To-Text Fired
> >  Time  0.200: Module :MOTOR running command MOVE-CURSOR
> >  Time  0.200: Module :VISION running command MOVE-ATTENTION
> >  Time  0.285: Module :VISION running command ENCODING-COMPLETE
> >  Time  0.285: Vision sees TEXT100
> >  Time  0.285: ?Text-Found Selected
> >  Time  0.335: ?Text-Found Fired
> >  ...
> >
>
>
>
> Without seeing the model I can't really say, but it definitely looks
> like there's either a bug in ACT-R or in your model.  If you send me
> the full model I can take a look at it and give you more details and
> fix the bug in ACT-R if that's what's wrong.
>
> Ignoring the problem with the backward time jump and just looking
> at this piece:
>
> >  Time  0.000: Vision found LOC103
> >  Time  0.000: ?Recall-Text-Location Selected
> >  Time  0.050: ?Recall-Text-Location Fired
> >  Time  2.050: ?Test-Recall-Not-Successful Selected
>
> I want to comment on your concern about the "time spent on
> ?Recall-Text-Location".  Conceptually, there was only 50 ms
> spent on ?Recall-Text-Location - from time 0 to time .05.  The
> two seconds from .05 until 2.05 when ?Test-Recall-Not-Successful
> was selected were spent "idle" in the procedural system.  Those
> two seconds were time that the declarative system spent trying
> to find a chunk as requested.  If there had been any productions
> that matched the contents of the buffers after ?Recall-Text-Location
> fired but before the declarative system signaled the failure
> they could have fired in that intervening 2 seconds.
>
> Dan
>
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: timebug.lisp
Type: application/applefile
Size: 856 bytes
Desc: not available
URL: <http://mailman.srv.cs.cmu.edu/pipermail/act-r-users/attachments/20041104/5d20a6c5/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: timebug.lisp
Type: application/text
Size: 2945 bytes
Desc: not available
URL: <http://mailman.srv.cs.cmu.edu/pipermail/act-r-users/attachments/20041104/5d20a6c5/attachment-0001.bin>
-------------- next part --------------