[quoting Davis, 11 Nov 95] Computers are a universal tool first built because they can do things easier, faster, and better than humans without the so-human mistakes that plague us (at least after the program has been debugged). I believe we should set our goal higher. We should build interactive learning systems that "do" learning better than we could ever do it! And why not?
How could that happen? Unless I've missed some major breakthroughs in artificial intelligence, our computers still can't contextualize, break out of the bounds of relevancy the programmer has set for them, etc.
The thing I liked about the virtual environment described in the paper (and I guess this could take place in other interactive platforms as well) is that it has room for that best part of interaction--the inclusion of learners and instructors in the microworld. Which, I suppose, is opposite of the point of the thing, which is to have the computer do the teaching--but I still think that some collaborative interaction would be valuable. Maybe this wouldn't be possible in a self-contained learning environment, but would it be against all the expectations of interactive multimedia to also let the student e-mail an expert, if needed?
I'm pretty sure I'm not explaining myself well, so let me ask this another way: in creating these environments, where computer-human communication seems key, would it not also be valuable to include the option for human-human communication?
I started wondering about this in a more serious way, having seen some examples of the kind of thing Harry describes below--Netscape, for example, pushing its ability to run Director presentations, etc. Obviously, you want the computer to handle just about every question the student will have--but, much like the cases Rod Sims outlines, where if you design the program to recognize any number of right answers, the student will inevitably come up with one more right answer that the program won't recognize, I can't help but imagine a real need for human-human interaction. If the program offers things other people have answered, in an attempt to give context to the student's own "wrong" (read: not programmed) answer, is that enough? I don't know; I'm just curious.
I wish to offer a challenge to this group. We talk a good game, but can we produce a practical application for all this theorizing?
For the sake of future paychecks, I hope so!
I propose that we INTERACTIVELY construct what we're talking about to serve as a model for others. How to do this? There is a Website set up by Paolo Tosolini that allows a person to distribute an interactive PC application (using MMToolBook). The remote user is able to launch it with a single mouse click from the Web page. Of course, there is some set up on the receivers' end, but nothing drastic.
Just a quick note on this part--I'm not altogether clear what you're suggesting, but what about all of us Mac users? Do you happen to know if there will be any applications that are multi-platform? (I'm completely ignorant on this, because so far all that Netscape has offered for this sort of thing is for PCs)
I propose that we build an interactive application with input from all, but updates to the Website by only one person (to avoid conflicts), using Rod's paper as a guide and putting together the "ultimate representational model." "Build it and they will come!"
By saying "ultimate representational model," what do you mean, exactly? Perhaps I misread the paper, but I didn't see a necessary valorization of the interactivities that were later in his list, in exclusion to the earlier (and probably much easier to program) ones.
I am no MMToolBook expert, but I think with all the brain power on the net we could manage nicely! I suggest this educational scenario: adult learners with prerequisite background that have to learn a particular piece of equipment: description, operation, use, and set up. Length of course is not a factor. Cost is not a factor. (I guess this will have to be a virtual project!). Platform is not an issue.
Again, I'm unclear on this. (Someone bop me on the head if I'm just being dense here.) Are you just talking about the design of an instruction for using this equipment? Or the actual implementation, using MMToolBook, of that instruction? And, what level might it take place on? After all, a plain ol' manual might work well for such an instruction, leaving out interactivity altogether.