Good Idea. And maybe divide the students into two representative (or random) groups (maybe that's what you had in mind):
Group 1) Learns Ada first, then Java.
Group 2) Learns Java first, then Ada.
This reduces the effect of depriving some students of Java or Ada - something to which opponents of an experiment might object 8^) This could be done to compare any two languages.
Try and keep it to a single course, switching mid-way through, so that the test group and instructors do not change during the experiment. Try and use an experienced Java instructor and Ada instructor. Then measure
* The time spent using the language (classroom hours)
* The number of concepts and depth of understanding (i.e. some function of aggregate grades on tests)
You would want to keep classroom hours constant and compare grades (*or* measure classroom hours needed to get to a certain grade.) One of the big variables which cannot be controlled will be drop-out rate. Can anyone suggest other variables which can be controlled (Steven has a list of other measures in his e-mail)?
This may also help transition the department in case the decision to go to Java wins out. (I am a fan of Ada and know very little Java - so I'm not advocating Java, here; just dealing with possibilities.)
Is there a web-site which reports the results of experiments in language comparison in the classroom?
[log in to unmask]
At 11:57 AM -0500 3/10/00, Steven Deller wrote:
>On Friday, March 10, 2000 7:02 AM, [log in to unmask]
>[SMTP:[log in to unmask]] wrote:
>> Also if anyone has taught CS1 using Ada with an OO approach from the
>> start, I would be interested in hearing of their experiences.
>Building on the reference by W. Wesley Groleau about John McCormick's
>experience with C and Ada, is there any possibility of doing a side by side
>comparison of Ada versus Java in an introductory course?
>That is, rather than fight the system, could you suggest that Aston
>University do some fundamental research into the comparative capabilities of
>the two languages vis-a-vis teaching OO programming. Are there enough
>sessions to do half in Ada and half in Java, and then compare somehow the
>student results from each (dropout rates, finished projects, student
>responses, subsequent grades in computer science courses, language
>independent "concepts grasped" exams)?
>I could easily argue that preparing an OO class to be taught in *both* Java
>and Ada would force the OO issues to become clearer, and would result in a
>better OO-concepts course than doing OO purely in only one or the other
>Steven Deller, Apex Ada Marketing
>[log in to unmask], (410) 757 6924
>Rational Software Corporation, http://www.rational.com
>For user email groups, check http://www.rational.com/support/newsgroup