Having been an instructor in the past, I am familiar with how difficult it can be to keep up with the daily grind of developing effective instruction. However, developing effective instruction starts with changing beliefs. Adding use of a few instructional principles here and there is not so difficult if the instructor trains himself to look for the opportunity.
Let me throw into the pit what I believe to be a very effective model of the learning process. I believe many of you are unfamiliar with it: The Franklin-Quest Reality Model. Anyone else familiar with this model? If not I suggest you research it. Learning is largely a matter of changing the beliefs on which our actions, cognitive or physical, are based. This model serves as a tool for changing beliefs and behaviors which serve as the evidences that learning takes place. Combined with other instructional models, this becomes a very effective tool, particularly on the personal level. Creative designers can discern how it applies in other situations as well.
While I agree with most of your comments listed below, you seemed to miss my unwritten main point: Example is one of the best instructors. I would propose that the more instructors involve students in their projects, the more likely that effective instruction would be taking place. Here are a few comments:
Serious ID isn't hard--its fun. Most effective designers like what they are doing. That's why they work effective design principles into their personal presentations. It's the ones that think of ID being hard that rely on telling. But serious ID is time consuming.
[quoting Foshay, 27 Feb 97.b] (3) The more you know about a subject, the more independent a learner you can be, and the less you need in the way of heavy-duty instructional support.
Statement #3 brings us back to reality. Most of the responsibility of the learning process at the University level belongs with the student. Effective instructors act as facilitators between students and knowledge.
Concerning "scientifically proven principles": Since it is true that there are few scientifically proven facts to rely on when it comes to instructional design (David Merrill may disagree), these become very important. Consider this pyramidal paradigm of instruction: Instruction ought to be based on scientifically proven facts. As we get further from the base, we move more towards the point, which represents pure speculation. In-between are theories and hypotheses based on varying degrees of scientific evidence. Most instruction should be designed near the base of the pyramid. Experimentation takes place nearer the tip of the pyramid.
Data collection, however, works in just the opposite direction, much like a funnel. The more speculative the instructional process, the more data must be collected, and the more scientifically-based the instruction, the less data must be collected, at least in terms of scientific evidence for the process. The funnel paradigm is used here because all data is funneled into developing more scientific evidence.