Jeffrey Elkner wrote: > On Sat, 22 Apr 2006 18:00:27 -0500, Michael Tobis <[EMAIL PROTECTED]> > wrote: >> I think I've got a realistic design for a doctest-driven beginners' >> programming courseware suitable for presenting HHTLCSpy - type >> materials that can be put together in a week or so with my level of >> skill. (Or a half-day if I let Ian B in on it... ) That week >> probably won't happen before early June, but I think I'll have >> something of immediate interest to you early enough in the summer >> that we can develop real curriculum in time for next fall. > > Thanks, Michael! I've written up a brief description of the tool I'm > looking for here: > > http://dc.ubuntu-us.org/projects/doctest-quiz.php > > and a briefer launchpad spec here: > > https://launchpad.net/products/cando/+spec/doctest-quiz > > Paul Carduner is interested in working on it, and I'm looking into > submitting it as a Summer of Code project. I'll keep you posted.
Have you guys contacted the people doing sandboxed Python on the web? I've seen two of them, but I can't for the life of me remember where and my searching has been fruitless so far. I think both people hang out on comp.lang.python. Of course, if it is running off localhost or otherwise not interactive-over-the-web then you can avoid the problem entirely. Though then you have to be sure you trust the doctest, since it could contain malicious code that would be run on the student's computer (it's more obvious if the code is malicious, but at least you want to be careful about it). If there were student-contributed doctests this seems like a potential concern. Of course if you were actually running a course, you could potentially just give each student a user account on a Unixy host and not worry about what they do. This would avoid the tedium of helping people get working and consistent Python development environments, and you could start adding things like version control to the base setup for each student. One of the things the doctest model might make possible is to avoid editing live code (which Python isn't good for). Since everything is repeatable you just start over each time. The tests would run in a separate process from the editing environment. Though you'd still want the test environment to hang around, so it'd be inspectable. But doctest's standard reports are enough to start with, so that feature could be added later. -- Ian Bicking | [EMAIL PROTECTED] | http://blog.ianbicking.org _______________________________________________ Edu-sig mailing list [email protected] http://mail.python.org/mailman/listinfo/edu-sig
