(Virtual) ITiCSE 2021 Reflection

I attended ITiCSE this year. One of my former undergraduate students, Anshul Shah, presented his paper on our CS1 reviewer app. Attending was an opportunity to iterate on how I approach online conferences and try out ideas from my last reflection when I attended SIGCSE 2021. One way to describe my approach to ITiCSE is less like prior online conferences and more like how I approach in-person conferences.

[This article is also posted on my personal blog so it is freely accessible.]

How I decided what to do

Step 1: Look at the proceedings two-ish weeks before and note all of the things I find interesting. — I did this based only on the titles. The abstracts hadn’t come out at the time.

Step 2: Go through the list a second time and mark which things I will actually attend. — Talks were live, and the videos would not be available until after the conference. I often had more papers I was interested in than I had time for, so I needed to prioritize. I purposely separated this step from Step 1 because I didn’t want to get bogged down during Step 1 debating what to go to, which is a different set of criteria than “is it interesting enough that I want to at least learn more.”

What I did

I blocked out the actual conference days as if I was going to be away and wouldn’t be able to do anything except conference stuff. So the most I was planning to do was email and my daily reading and writing habits. I also gave myself permission (a.k.a. to not feel guilty) to not do my daily reading and writing goal if time got away from me.

I wasn’t perfect, though. I scheduled a few meetings outside the conference hours that were still during my business hours. The conference was thankfully during reasonable hours for my time zone, 8 am to 2 pm. I just had to skip my morning walk and get up a tiny bit earlier. The meetings were a mix of the usual weekly ones like my research meeting with my summer undergraduates and one-offs that would have happened at the conference if it was in-person.

I made social opportunities a higher priority than talks. This priority was even though it meant I wouldn’t see the missed talks until ITiCSE posted the recording after the conference. This priority also meant the opportunity cost of not following up with an author if I learned something compelling during the talk. However, in my mind, I’m more likely to find people I resonate with at social events than the process of “attending a talk, find it interesting, and try to connect with the author(s).” I suspect that for others, your mileage may vary. Moreover, I’m more looking for colleagues to discuss and critique ideas with than new projects.

This social event priority meant I attended the Lean Coffees and skipped or missed the beginning of talks a few times. I absolutely loved the Lean Coffees. I’ve been raving about them at multiple venues. And I want to use the structure in other social contexts where a little structure would help make the conversation flow AND increase the odds the collective group gets something out of the conversation. Here are the steps as best as I remember them:

1. Everyone proposes as many ideas as they have on what to talk about. They had a specific tool that I no longer remember. But I think this would be possible on Padlet, except on Padlet the host can’t limit the number of votes per person.
2. Everyone has three votes and votes for the things they most want to discuss. If you want to restrict the number of votes, you could try using OpaVote.
3. The person that proposed the idea with the most votes gets five minutes to elaborate.
4. We spend another three minutes talking about it.
5. We then vote if we want to continue the topic or move on.
6. Repeat from Step 3

Conference Critique

I’m going to approach my critique the same I would a new way of teaching. Online conferences are new, and we are still figuring things out. Therefore, we should view everything we do as an experiment. Experiments are allowed to fail. What’s important is to learn from the experience and articulate the experiment’s goals to see what worked and didn’t work.

What went well

As I said earlier, I loved the Lean Coffees and wished there were more. Using Moodle as a conference platform was fine. In some ways, it was odd, but in others, it was not a big deal despite it not being a dedicated conference platform. I think that experiment worked out. I appreciated that the entire program was on a single page with all the necessary links and information. I also liked the ohyay instance for socializing.

What could have been better

I wish more people went to the ohyay, but I suspect part of it was that being online makes it much easier to “leave” the conference context. That ease is good in some ways, but it puts more responsibility on the participant to decide when and how to stay in the conference context. I’m not sure if there is any way for the conference organizers to change this, except for brainstorming ways to reduce the friction to socialize and make it more attractive.

I really want videos. If pre-recording is possible, I think it should be, if at all possible. Bonus if it’s released as early as is reasonable. When I asked why we were required to present live, the rationale was to allow the talk to be less artificial and more organic, such as connecting their talk to what was happening in the conference.

I would like to respectfully disagree with this rationale, and I’m going to draw on the rationales we’ve been using when discussing how to make the learning experience better for our students during emergency remote teaching. Let’s frame this in terms of benefits versus costs/harms.

I can see that requiring live presentation enables speakers to make their talk more organic and artificial. I can imagine this then makes the talk more engaging to the audience. And the cost is the same as it was pre-pandemic when we were all going to conferences in person. However, this plan is the same as deciding for emergency remote teaching to simply lecture over a video meeting, copy the in-person world and paste it into the virtual world without taking advantage of the virtual space. I hope most of the audience of this blog recognized if/when they did this in their teaching that it was not the same and insufficient for their students.

The benefits of videos include asynchronous consumption, ensuring participants hear the talk from the beginning, skipping around the video, rewatching sections, and catching up on the talk by watching it at a faster speed before Q/A happens. And this list is only the benefits that I have experienced! I’m sure there are more. Another benefit is that the speaker now has a video to engage potential consumers of their work (because papers can be hard to read at times, especially for those new to the field.). The cost is asking speakers to create a video on top of their talk, and earlier than a talk would need to be ready for proper video-captioning. This request requires new skills and, therefore, will take extra time. But the time to learn the skill is a one-time cost that they may have already paid or can reuse in the future (because videos are here to stay, and I’m glad of it). Once they pay that one-time cost, the direct cost is extra time recording and producing the video. Some of that time can be recouped by not having to practice the talk until the speaker has it perfect because editing is a thing of wonder.

So given the benefits and costs of live online talks versus videos at a conference, I’m on the side of videos. And in the future, when I give a talk, I plan to do what Amy Ko is doing and release a video of the talk. I think the benefit is worth the cost. I think I’ll even put a link to the video and slides on the bottom of my slides if I give the talk live so those that missed something can go and catch up there.

A counterargument to the above is I can’t claim that emergency remote teaching is the same as emergency remote conferencing. While true, I do not think this is a strong argument given the goals and context. The conveyors of knowledge (teachers vs. speakers) and participants (students vs. conference-goers) are different. The interaction model, goals, self-regulated learning skills, and preparedness of participants are different. However, within the context of a conference talk, the goal I argue is pretty similar. Knowledge transfer between the holder of knowledge to the participants. Also, within the conference context, there are other goals like talking to the speaker for further questions or potential collaborations, discussing with peers in the audience, etc. But a live talk will not help those goals and may harm them. I’ve been to plenty of live and online talks where I missed the beginning, got distracted in the middle, or the speaker refers to another talk I wasn’t present for, so I didn’t understand the reference. I was then lost for the rest of the talk.

On the other hand, I am not saying that we should expect a talk’s audience to watch the video in advance and therefore show up to the talk prepared to immediately start Q/A. Learning@Scale tried that a few years ago when they attempted to flip the conference. It was an experiment worth trying, and we learned that it doesn’t work. Instead, we should do the equivalent of a silent meeting, where we play the video first and then Q/A. This process is what ICER 2020 did, and I loved it. They are also doing it in ICER 2021. Bonuses are Q/A happens while the video plays, and all the authors help answer questions rather than focusing on the single speaker.

Summary

I think I’m getting better at online conferences! Though, it helped a lot that the hours were within my regular work hours. I’m a little fearful that as we transition back to in-person, we will lose all the benefits to the online part. So many people seem to have rose-colored glasses when it comes to pre-pandemic conferences.

I know this blog post’s “What could have been better” is way longer than the rest of it. It was an experiment worth doing, and it did not work for me. What do you all think? Did I miss the mark on my critique? Is there a cost or benefit I missed when it comes to talks?