About mandatory attendance

19 Oct 2022

There has been a lot of discussion in the CS research community about the “conference culture”: the fact that the main publications in our field are conference publications, and its consequences.

In this post I would like to discuss just one bad aspect: traveling to conferences is bad for the climate. I propose to consider one modification to the system, which is to remove the rule that one has to attend the conference to have the paper in the proceedings. I will summarize quickly the motivation and proposal, and then try to discuss the issues of this proposition.

The carbon-recognition correlation

Prestigious publications are crucial to us: it is the way to get a permanent job, to get funding, and to be recognized by the peers. In CS, the prestigious venues are the international conferences, and attending international conferences very often involves flying. In other words, our system creates a correlation between recognition of our work and carbon emissions.

It is safe to say that this is a wrong incentive, given the current climate crisis.

Before going further, let me comment on the answer that would be “yes, it is bad for the planet, but our job is important, others should make the effort, not us”. One could argue about how solid this opinion is, but instead I’d like to make two points. First, not everybody in the community thinks the same way, and we should try to create space for these other people. Second, it is very improbable that in 20 years the conference system will still work as it does, so it is good to start changing some things now.

What to change?

There has been intense discussion about what to change in the system: moving to a journal system (like in maths), stopping some conferences, etc. These are important discussions, but the proposed changes are big steps, and as a community it is difficult to take such steps. A friendly atmosphere is somehow needed to keep doing reasonably good and fair science, and bold moves create animosity.

A small step I propose to consider is removing this one sentence that every CFP has: “Any submission accepted into the technical program but not presented on-site will be withdrawn from the final proceedings” (this one is from DISC 2022 CFP). In other words, I propose to abolish mandatory attendance, and to allow people who do not want to attend physically to have recorded talks or online talks.

In this post, I don’t want to discuss what would be the exact replacement, which kind of hybrid conference would be the best, etc. Also, since this is a small step, that could be reverted or be followed by other changes, I don’t want to discuss what would be the dynamic in the long run. Let’s say that we discuss what would happen if we do this for five years.

Critics and answers

Below I have listed some critics of this proposition, and tried to answer them.

The travel-publication correlation is good for some people

I have used the expression “carbon-recognition correlation” to highlight the bad aspects of the current system, but for some people the correlation between having good publications and traveling to conferences is actually a good one. The typical reasoning would be: “For my work I need to meet my colleagues regularly. My institution/grant would pay for traveling only if it yields to more prestige. I can trade high-level publications for travel funding thanks to mandatory attendance.”

This is a real phenomenon, but I would like to discuss two aspects. First, it does not affect every researcher. Some have enough money anyway, and can for example go to workshops that do not yield to publications. Some would not get any funding anyway, because they don’t have grants, and their institutions would not fund the trip to a conference. Only the people in between would be hurt by the change. Second, it is very bad if these people are indeed hurt. But, note that with the current system there are already people who want to attend and cannot. My opinion is that we have to tackle this problem in a more satisfactory way than by a kind of extortion. Maybe some travel grants, maybe more local events, etc.

Mandatory attendance is good for organizers

As a local organizer I see that mandatory attendance is very reassuring: you know that approximately X papers will be accepted, thus there will be at least X people, and this is useful when balancing the budget.

I would answer that: (1) anyway, there is always some fraction of the participants who do not present a paper (so X is just a lower bound), (2) after the first year without mandatory attendance, one would already have some statistics, (3) workshops (even large ones in maths) have done this for years, and (4) asking early for registration or some kind of soft commitment could help.

Of course, it is very difficult to be a local organizer, and it is pity to make the job even more difficult, but I think it is worth it.

Hybrid conferences have issues

I think everybody now knows that hybrid conferences are sometimes not great. In the following I’ll take the perspective of someone who would attend onsite, since people who attend online would have made this choice, knowing the difficulties.

Problem 1: Recorded/online talks are not as good as real talks. This is quite true, partly because of technical problems, partly because seeing people for real is a better experience. But I would argue that anyway, conference talks in general are not very good, and that I prefer a good recorded talk to a bad live talk. That is, there is a much larger difference between the talks themselves than between “technologies”, so it is more important to optimize the first, not the second. Also note that recorded talks could go through a round of reviews+revision, which could improve the quality of lot.

Problem 2: It is not convenient to communicate with people that are not on-site. Again, I partly agree. For the questions at the end of talks, I think that anyway, they are not always useful (either too specific, or too general, or just odd). For the little chat at the coffee break, at the restaurants, breakfast, yes indeed, we are loosing quite a lot with hybrid. Beyond just the fun of it, these discussions build trust, which is essential when one needs someone for advice, for reviews, for collaboration, and more generally to build a community. I hope this trust can be created a bit differently, via internet, or with more local people, or during fewer events that could maybe be longer, etc. Again, other communities do not use so many international conferences, and they seem to work pretty well.

More comments

I got some feedback on the first version of this post, and here are two things I forgot to mention:

• One way to relax mandatory attendance without completely removing it is to require that people present their papers on-site at some point. For example it could be in the 3 years after publication, and if there are several conferences of similar strength in the field (eg DISC/PODC), one could maybe choose which one to present at. Then one could choose the conference place that is closer to her/his place. We could call this delayed mandatory attendance.

• There is one aspect that is not related to climate change, but has to be mentioned to balance the drawbacks of hybrid I mentioned: for some people it is very difficult to get visas to go to conferences, thus having something online is much better for them.

(Thanks to Joel and Darya, for pointing this out.)