Software Carpentry at the AAAS

I just received word that I’ll be running a workshop on the aims, benefits, and curriculum of the Software Carpentry course at the Annual Meeting of the American Association for the Advancement of Science in St Louis on Friday, February 17. The AAAS AM is the biggest gathering of scientists in the world; I’m pretty excited.

Meanwhile, Michael Hoffman is teaching a short course on Python at the European Bioinformatics Institute in Cambridge, England, based on the course notes. 24 people are now enrolled in the course at the University of Toronto, 47 are auditing locally, and 21 others are sitting in from other locations. (These number don’t include the students enrolled in the course at Indiana University, or the study group at CalTech.) Oh, and 38 students have already submitted Exercise 1, even though it isn’t due until 5:00 Friday ;-).

Day 9

Week two of Software Carpentry, and things are starting to settle down. There are now 93 (!) people signed up for the course:

	U of T	Elsewhere
Auditing	18	21
Enrolled	25
Unknown	29

Of those actually at the University of Toronto (as opposed to the
local hospitals—a few radiologists are sitting in) the breakdown
is:

16	Computer Science
14	Physics
9	Civil Engineering
	Mechanical and Industrial Engineering
5	Biochemistry
	Institute for Aerospace Studies
2	Institute of Medical Science
	Mathematics
1	Astronomy
	Biomaterials and Biomedical Engineering
	Botany
	Geology
	Medical Biophysics
	Zoology
	Non-degree student

This week’s lecture went much more smoothly than last week’s, in
part because we were in a larger room, with seating for everyone, and
in part because the content was an introduction to Python, which I’ve
given more times than I can count. There were still some glitches in
the slides, though: a few things were out of order, and I really do
need to choose more concrete examples. 33 students have completed the
first exercise, which is already more than I’d expected—I’m feeling
uncharacteristically optimistic right now 😉

The past week has also CSC49X, the fourth-year Computer Science project course. 23 students are working on 10 different projects (24 and 11 respectively, if you count
Sean Dawson’s paid-but-not-for-course-credit work on a sequence diagram debugger plugin for Eclipse). DrProject (our Trac-derived project management system) is holding up so far, and we’re almost finished setting it up for two other courses to use this term as well. Not bad for Day 9 of term; not bad at all.

Software Carpentry: First Meeting

The Toronto edition of the Software Carpentry course met for the first time on Monday. 51 people crammed into a room with seating for 34 (and no air conditioning, on an unseasonably warm day): medical biophysicists, computer scientists, civil engineers, and even a couple of faculty members.

I was pleased with the turnout, but less pleased with my lectures. I don’t think my introduction to the shell made sense to anyone who didn’t already know the content—it was far too long, and as Andy Lumsdaine (who’s teaching from these same notes at Indiana) said, this stuff really does need to be interactive.

The version control lecture was more successful. While the notes talk about editing C files, I talked about co-authoring a paper written in LaTeX, which I hope was less intimidating. (A couple of people have suggested that there ought to be an entire lecture on LaTeX, but I’m still not convinced—it would be a good sample problem for the lecture on Make, but it isn’t really software engineering itself.)

I spent several hours yesterday (Tuesday) editing the recordings I made of my lectures using Audacity, then converted them to MP3’s with CDex. I have a few verbal tics, and we spent more time shuffling chairs around than I thought, so each of the lectures reduced to under 30 minutes of real content. It could be that the biggest thing I’ll get out of this course personally is better public speaking skills… 😉

The rest of yesterday was spent putting together some Python scripts to manage the class list, and to generate Subversion passwords and access control file entries from it. I also played with two GUI interfaces for Subversion (RapidSVN, which is cross-platform, but not really very rapid, and TortoiseSVN, which is Windows-only). I need to check out SmartSVN as well, so that Macintoids will have something to play with—it’s clear Monday’s lecture that unless they can start with a GUI, many students will go away thinking that version control is intrinsically hard.

With another hour’s work, I’d be ready to send mail to students in Toronto and elsewhere telling them how to access the Subversion repositories they’ll be using in the course, and what their first exercise is. However, this morning is going to be taken up with the first batch o 49X project meetings. This term, 23 students will be working on:

• a collaborative grading tool;
• a combined land use and vehicle traffic simulator;
• neuroimaging algorithm performance;
• an on-line marking aid;
• visualizing rock planes in 3D;
• a new interface for the Bell Kids Help Phone;
• a lightweight requirements management tool;
• DrProject itself;
• a database for a food bank; and
• an electronic GFP (Green Fluorescence Protein) browser.

It’s going to be a busy, but rewarding, fall…