Towards immediate in-context feedback Paul Calder Computer Science, Engineering, & Maths Automated Feedback
Flinders University / Computer Science, Engineering, and Mathematics 2 Automatic Feedback Students Timely: as soon as they have something worth trying Frequent: encourages continual improvement No surprises: known to satisfy requirements Teachers Assessment: testing functional performance is (almost) free Coverage: thoughtful test cases ensure full conformance Scaffolding: basic capabilities before advanced Consulting: students come with more focussed questions
Flinders University / Computer Science, Engineering, and Mathematics 3 Where We Are Now Capabilities Computer programming labs, tutes, assignments Functional specification conformance Progressive tasks: levels of performance Limitations Only on campus: CSEM lab opening hours Cumbersome interface Feedback is not in context Not integrated with submission (or assessment)
Flinders University / Computer Science, Engineering, and Mathematics 4 Where were heading Immediate goals Wherever and whenever the student is working: home, on-line,... Embedded into authoring tools: feedback in context Non-functional feedback: it works, but is it good? Integrate with electronic submission: Moodle plugin Longer term FLO: need thorough testing first! Other easy domains: engineering, mathematics,... More challenging: physical sciences, economics,... Aspects of many domains: Clippy done right?