The paradigm underlying the thesis here, adopted by others in the discussion, is that of the machine as GOLEM--human-like in many aspects, with the threat of usurping or replacing us in one way or another. A different view of the machine is that of enhancer, or potentiator, which makes the human more than we could be without the machine. The machine as mechanical and process-enhancer is commonplace; however, the computer as a cognitive enhancer is not.
Instead, then, of Alessi's view that machines that read and write will do so for the human cognitive potato, consider that in five to ten years we could have computers that are so effective in enhancing human cognitive and learning powers that anyone can learn, with such a machine, to read and write fluently.
The discussion has taken the direction of: well, machines that read and write are not in the future, they're here now, and what is the nature of learning going to be like if machines read, and what is learning-by-hearing like compared to learning-by-reading. This assumption that if we can do less, we will needs to be challenged.
What, you might ask, is the likelihood in five to ten years of mind-enhancing machines that will facilitate the increase of human cognitive powers? In 1983, the research facility that I started in Computer Science at the University of Victoria was called The Laboratory for Computer Enhanced Cognition. Alex Pentland at MIT's Media Lab has stated that his research goals are in the uses of the computer to enhance human capability. I continue to work in this area, through the creation of interactive learning systems based on this paradigm shift.
It seems to me that, following the notion that memes (ideas as viruses) guarantee that if you can think of anything at all, someone else will have, or is right now, or soon will be, thinking of the same idea independently, that computer enhanced cognition is the learning and education research opportunity emergent. So no doubt, there are many others whose work will produce those real gains in human cognitive capability that has been so elusive.
The argument that Steve puts forward seems to consider individual learning and collective learning as being in opposition, as approaches from which we must choose one or the other. He then goes on to state that software does not support collaborative learning very well (as yet), and in fact, that educational technologies develop systems which are the antithesis of the kind of interaction one might want in collaborative learning.
However, these two concepts need not be seen as a binary polar model; they are, in my opinion, orthogonal. First, for clarification, let's take "individual" learning in the sense of "being by myself" learning. The best kinds of things to learn, being by myself, are the base concepts and relationships of a subject: kinship models in anthropology, Newton's laws in physics, etc. Collaborative learning in the sense of "being with others" is orthogonal to this in the sense that what others contribute is the exchange of perspectives on applications of concepts, relative value systems, which enrich the individual's knowledge as well as offering opportunities to sort out contusions or misunderstandings.
From this point of view, "being by myself" learning deals with the "fact base" (risky as this term is in a deconstructivist age!), while "being with a group" deals with communications and value exchanges. The appropriate computer tools MUST be different, the first being task or information based, and the second being message based.
However, the two learning approaches are orthogonal and complementary in that having one does not include or preclude having the other. In fact, strong arguments can be made for having both.
There is another sense of "individual" learning--the knowledge gained by a specific person, and its corollary of "collaborative" learning--the knowledge gained by the group. As Merrill stated, only one person is learning, as far as that individual is concerned, and that is true however many persons are present. The other, the knowledge of a group, is more controversial--sociologists measure and consider groups as entities, with their own specific behaviors, etc. This dimension of person-specific vs. group-specific is more troubling: in many group situations, the teacher, facilitator, instructor, deals with the group as an entity, but its knowledge or behavior is not that of any specific individual. Here, the person can become lost or missed by the facilitator because the "group" seems to be doing so well--20% of the individuals asking and answering all the questions.
This dimension has its own research questions. If a group seems to be learning, is the non-participating individual also learning (like secondary smoke effects)? For myself, I find the first set of meanings more compelling--always dealing with individuals, in being-alone or being-with-others situations. This to me is the proper domain of the use of the computer--in ensuring that the individual has a personal adaptive environment that facilitates individual learning and information needs.
Steve's relativistic acceptance of all instructional design models, and the creativity and artistry needed to execute them, leads to his "need for authoring tools which truly foster evaluation and revision processes."
This is either a domain in its infancy, or one in which there is no (and may never have) guideline for standards and value. Every process is a creative one, and every product is suspect and needs to be tested before one can know if it is effective! I'm not against evaluation, but the effort of thorough evaluation of every product is not practical or affordable.
Guidelines for building bridges or designing with transistors are based on established principles, and building within those guidelines is less art and creativity than following rules (and malpractive if you don't). There is an element of predictability in terms of process and outcome.
Art, on the other hand, is a process which is creative, for which there are no well-established guidelines for what is "good" or "works." The leaders of the day direct trends, and they can be deconstructed. There is no clear basis for evaluation.
Instructional Design may be art, but accepting that it is so is anathema to those in the system who need to develop or convert the tremendous backlog of traditional material into more effective learning products.
If it is not art, then we had better find those standard processes and outcome predictors, even at the cost of converting artists into learning-engineers.