Thursday, March 7, 2013

SIGCSE 2013 Conference Report, Day 1

The opening keynote was unconventional: rather than have a single distinguished speaker, they had seven people give "Flash Talks," with twenty slides, fifteen seconds per slide, all on the theme of broadening participation. It was kind of fun, but if you follow broadening participation at all, then you likely knew everything they had to say. The nice thing about a Flash Talk though is that even if you don't care for the content or speaker, you don't have to wait long for the next one. I've still never done one, but it sounds like a good design challenge. I do wonder why they didn't do a proper pechakucha.

One of the presenters was Jan Cuny who spoke about the NSF's CS10K program, which seeks to get 10,000 qualified high school computer science teachers by 2015. Two things she mentioned interested me: that high school was where intervention dropped off, and that there was a plan. This inspired me to ask a question about the role of K-8 in the plan: if a diverse group of students didn't consider computing as a career, who would fill the high school classes? Her response was that the NSF is currently focused on high school precisely because the culture of AP courses provides a hook: there's a place for an intervention to live where, once established, it is not as likely to be a victim of cutbacks in the way that a purely elective computer science course might. She (or another respondent, I don't remember) also mentioned that the K-8 interventions are primarily about computing in a context as opposed to Computer Science as a distinct field. This makes sense, since few K-8 schools have a CS teacher but they all have to do math, science, English, etc. It's true that we will lose inspired K-8 students if they have no where to go, and I will be interested to track this project as it goes on. I'm sure future SIGCSE meetings will report on it, if nothing else.

I went to several paper presentations, and while they were interesting, I don't have much specific to share from them. While the SIGCSE community has moved toward a more rigorous research model, there is still a lot of stuff that gets published that strikes me as immature work. I think a good path to raising the overall level of quality would be to accept fewer, high-quality works, and have more attendees in the room seeing what good work looks like. Right now, with four or five concurrent paper sessions, attendees are spread around and subjected to a mix of excellent, good, and fair work; a novice in the field would not be adequately prepared to tell the difference.

Particularly troublesome to me—and I've mentioned this in my other conference reports, so you know it's coming—is the low quality of the presentations themselves. Every presentation had bullet points on slides with lots of text. Several had unreadable visuals, either due to poor design or poor scaling. Yet, these presenters are all Computer Science teachers who ostensibly study education: why are their presentations so poorly designed? The most ironic one was the presenter who pointed out that people say undergraduates don't know how to design programs, and that undergraduates don't know how to design algorithms, ... then he proceeded to demonstrate that he didn't know how to design a presentation.

At least twice today, I saw speakers get tripped up by attendees asking two-part questions. I need to remember to only ask one question at a time. My talks are designed to allow for some short discussion afterward, and I may go so far as to interrupt anyone who says "I have two questions..." and demand only one at a time, to ensure quality of conversation. I know I saw a speaker do that once, though I cannot remember where or when.

I got to see the presentation of the Best Paper Award winner, as well as related work in the same session. Briefly, this was the work that Mark Guzdial wrote about, how at UCSD they combined media computing, peer instruction, and pair programming to dramatically increase retention and success. It's very cool work, and I'll probably have more to say about it later. Long story short, their results are too good for us not to try to replicate them.

In the exhibit hall, I talked to a representative from GitHub and got to ask him about their academic agreements. I might consider trying GitHub in a future project. I originally chose Mercurial years ago because the command line struck me as more learnable for novices than Git's. Turns out Git has unified into one executable and basically copied Mercurial's syntax (from what I can tell), so it's back on the table. I also spoke at length with two folks at the poster session, doctoral students who were both presenting work that was not really their dissertation work, both related to games and learning. I'll be sure to keep an eye on both projects, as they coincide with some ideas I would like to pursue in the short term. I've found that I really enjoy the poster session format and talking with young scholars. Maybe it's because they are more willing to receive criticism and feedback than some old scholars. Memo to myself: remember to accept criticism and feedback.

I have been able to meet with some old friends and catch up, which is always nice. I remember the first time I went to an academic conference, it was very intimidating to know nobody. It feels great to see people at SIGCSE who I think are interesting and who are doing good work.

In other news, I had a medium rare burger that was actually medium rare along with a local lager at Katie Mullins last night. Today, I had a bowl of pho with a friend for lunch today—a welcome treat since although Muncie has recently acquired some great ethnic restaurants, we're still lacking Vietnamese. Denver is not what I expected: never having been here, I expected it to be more mountainous. In fact, the drive from the ~30 minute drive from the airport to downtown was very flat. Looking one direction, you'd swear you were in Kansas. The other way, behind the city, beautiful mountains. Maybe this is old news to folks who have been here, but I thought it would be worth documenting my assumptions so that I remember to watch out for preconceived notions in more serious matters.

Next up: the reception, which is promised to feature a free drink and heavy appetizers. I'm looking forward to my talk tomorrow and just did some minor revisions on the slides. After looking at the schedule, I'm a little miffed that Saturday's talk is in the "capstone" paper session, since my work has nothing to do with capstones per se. I hope I get the right kind of audience to share this work and get good feedback.

No comments:

Post a Comment