[quoting Foshay's paper] Meanwhile, the truly unique requirements of TBT aren't being addressed, such as answer analysis (none of the present tools have as much power as the 30 year old, ugly approach of Tutor)...
This is the first time I've seen in print someone knock present authoring systems for their lack of response analysis handling. More often, old Plato gets knocked because one had to PROGRAM the whole lesson. Most people tend to forget that the most successful CAL courseware was the LDE Chemistry course developed for mainframe Plato. The PC port of this software, still unchanged after 15 years (heaven forbid that they do change it), is still marketed by TRO as part of their Plato system. It may look dated, but it works.
I have spent a good deal of time trying to simulate some of Tutor's response handling techniques in modern PC languages, and only now with the RAD systems available, which have a solid, functional programming language behind them, do I feel that I am winning.
Secondly, TBT and ID practitioners, in my view, never seem to take into account the most important part of the equation--the end user. They all feel they know what is best for the learner, but no one I know does follow up work to see how the learner interacts with and reacts to the application. In my situation, I work directly with the students and can tell very quickly, from their reaction, the effectiveness of the software. Unfortunately, decision makers who eventually buy the software are easily influenced by the "Oh wow!!" factor which is often found in lessons based on a questionable ID footing. The attitude of the purchaser is then: "Look at the millions we've spent on this new TBT, why aren't you using it?" The answer is simple, the people who are supposed to be benefiting from this software, have used it once or twice, become bored or disillusioned with it, and then refuse to use it.