[ACT-R-users] ACT-R & E-Prime

[db30 at andrew.cmu.edu]

--On Wednesday, December 01, 2010 4:59 PM +0000 j.a.grange at psy.keele.ac.uk wrote:

> Dear community,
>
> Is there any way to get ACT-R models to interact with windows other than
> those generated by a LISP-coded experiment? For example, is there any way
> to program an experiment in other software (i.e. E-Prime)and have the
> model interact with this?
>

Yes, it is possible to have a model interact with something other than the
experiment tools provided.  However, there are no alternative interfaces
built into the ACT-R software so one would have to create such an interface
for the model as needed.

At a low level, the perceptual and motor modules of ACT-R interact with
the world through what is called a device.  The device is responsible for
providing the visual percepts to the model and handling the motor and speech
output which the model generates.  Thus to create an interface to a
different environment requires creating a device for the model.  There is a
set of slides titled "extending-actr" in the docs directory of the ACT-R 6.0
distribution which describes what is necessary to create a new device, and
there are examples which go along with those slides found in the examples
directory of the distribution.

If that alternate world is also written in Lisp (as is the case for the
example devices) then writing the device is essentially all that's
necessary.  However, if the external environment is outside of Lisp then
there are other software issues involved which are beyond the scope of the
ACT-R software.  In particular, how to interface the device for the model to
that external software depends on what sort of interface that external
software provides and whether or not the Lisp being used has the necessary
tools for connecting to and working with such an interface.  Those sorts of
things are going to require consulting other appropriate sources.

Although there aren't any alternate interfaces included with the ACT-R
software, I have seen announcements by people with claims about general
interfaces for ACT-R over the years.  Here are a few which I remember [I
apologize to anyone who may have written something which I missed]:

SegMan by Rob St. Amant <http://acs.ist.psu.edu/papers/stamantRRR05.pdf>
ACT-CV by Marc Halbrügge <http://act-cv.sourceforge.net/>
An ACT-R Java interface by Philippe Büttner 
<http://www.zmms.tu-berlin.de/kogmod/tools/hello-java.html>

I have not tried any of them, so I can't say anything about their usefulness
or applicability to your current needs.


I hope that information helps, and if you have any more questions about
ACT-R feel free to ask.

Dan