Note: Flash and Java interfaces are still supported but may not be supported in future releases. We highly recommend implementing your tutor interface using the new HTML components.
| Feature | HTML | Java | Flash | Benefit |
|---|---|---|---|---|
| Example-Tracing Tutors - Create problem-specific tutors by demonstration, without programming. |  |
|
|
Can be built quickly; no programming |
| Model-Tracing (Cognitive) Tutors - Create a generalizable cognitive model of student problem solving by writing production rules. | |
|
|
|
| Drag-and-drop user interface building - Build tutor interfaces through a graphical program. | |
|
|
Visually assemble your tutor without programming. |
| Logging over the Web - Log student activities during problem solving. | |
|
|
Log to the PSLC DataShop; obtain data for running analyses. |
| Logging to disk - Log student activities during problem solving. | |
|
Log to a file that can later be imported to the PSLC DataShop. Flash tutors cannot log to the disk of the computer running the interface, but when using the Tutoring Service, the Tutoring Service can log to the disk of the machine on which it is running. | |
| Web delivery with TutorShop - Deliver CTAT tutors in sequence over the web; convenient for running online experiments. | |
|
In addition to TutorShop, HTML tutors can be imported into popular learning systems, such as Moodle, Blackboard, OpenEdx, OLI and any other site supporting LTI. | |
| Dynamic interfaces - Add steps in a behavior graph that the tutor will perform, which can modify the interface at runtime. | |
|
|
Better support in HTML and Flash, which allow the hiding and showing of many interface widgets at once. |
| Suppress student feedback - Run a tutor without any feedback to the student, allowing you to log student actions for a pre or posttest. | |
|
|
|
| Simulated Student - a machine-learning agent that facilitates authoring of Cognitive Tutors for educators, who are not AI programmers. Instead of building a cognitive model representing a task to be taught by programming production rules, you let SimStudent observe you performing the target task. SimStudent then induces production rules that replicate (and generalize from) your performance. | |