An open letter from scientists during COP28

Written by Antoine Amarilli
Classified in : Uncategorized Tags : none

As COP28 is currently taking place in the United Arab Emirates, we relay the open call of Scientist rebellion. Over a thousand scientists, including some TCS4F members, are warning the world about the impending climate disaster and the dire insufficiency of political action over the last 30 years. This letter has just been covered in the Guardian, and it is still possible to sign it.

An Ode to Train Travel

Written by TijndeVos
Classified in : Uncategorized Tags : none
As a pledger to the TCS4F manifesto, I try to limit the amount of flying I do for professional reasons. As a PhD student, making new connections is important to me. In the last few years, I've tried to attend important conferences whenever they were a train journey away - also when I did not have a paper there. This led to many good experiences, since train travel within Europe is easy and comfortable. 
However, to get your papers published at a conference, the sad truth is that almost all conferences still require in person attendance. Last summer, I had two brief announcements accepted at the Symposium on Principles of Distributed Computing (PODC), held as part of the Federated Computing Research Conference (FCRC) 2023 in Orlando. The FCRC is a mega event, where multiple computer science conferences come together, including for example STOC. It would definitely be interesting to attend, would it no be for the fact that I live in Salzburg (Austria), so it's kind of tricky to get there without flying. So although I was happy that my work was accepted, I was reluctant to travel there. After contacting the chair of the Organizing Committee, the Steering Committee, and the Program Committee, I received a definite "No" on remote presentations/videos for the sake of our climate. I gave in and started to plan my trip. 
 
PODC 2023 took place in the third week of June. Earlier that month, June 6 till 9, I was going to attend the International Colloquium on Structural Information and Communication Complexity (SIROCCO) 2023, held in Alcalá de Henares near Madrid (Spain). The reason to go there was that I had a single-author paper at SIROCCO, and they have a mandatory in-person presentation policy as well. I decided to link both trips to minimize my travel. Here's the story of how I attended, two conferences, one workshop, and visited two non-academic friends along the way. This took place in a little over two weeks, spanning 5 cities in two continents. And all of that with only two flights. And not two "trips including flying", but two direct flights. Oh, and a lot of trains. That is 13 trains with a total of more than 5000 km of train tracks.  

Getting to Spain

The first leg of my journey would be travelling to Madrid. I'd visit a friend there the weekend before SIROCCO. Madrid - Salzburg is only 1500 km apart in a straight line, but this line crosses the Alps and the Pyrenees. The fastest train line is almost twice as long, going over Paris. Perhaps a shorter train connection exists, but it wouldn't be faster. Besides, taking dozens of slowly moving mountain trains did not sound very appealing. I opted for the long high-speed route, which takes roughly 36 hours, including transfer times. 



This 36 hours might sound daunting at first, but let me split it out for you. On Thursday morning I took the train from Salzburg to Paris. In the train I finished a review that was almost due, making the hours well spent. For who knows Paris: if you need to transfer in Paris, you also need to change train stations. This is usually a very short metro ride, but I took the opportunity to stretch my legs. I took the 30 minute-walk to my favorite vegan restaurant in Paris: Veget’halles. After enjoying a nice dinner there, I walked another 30 minutes to the next train station. From there I took a night train to Toulouse. One of the advantages of night trains is that you arrive rested at the next city, often even after having had breakfast in the train.  

Since I had a long transfer (5 hours) in Toulouse, I took another break from trains and stations. For who knows me, it’s not hard to guess where I went. I spend the morning bouldering at bouldering gym Arkose. After a good workout and a shower, I headed towards the city for some lunch at a Chinese restaurant. The afternoon was – of course – spent in trains again. After transfers in Narbonne and Barcelona, I was in my final train to Madrid. I can especially recommend the Narbonne-Barcelona train. This one has beautiful views of the countryside, parts of the Pyrenees and the Mediterranean. I’d say there are worse places to work on your conference presentations. On the topic of working on the train: most European trains have good Wi-Fi nowadays. Even videocalls aren’t out of the question. However, this is hard to estimate if you haven’t taken a particular train before. I always make sure to have offline options, such as printed papers to read and LaTeX running on my computer rather than Overleaf. And as we all experience sometimes: being offline is often more productive than when you have access to the distractions of the internet.  

Train travel in the USA

I was very lucky that there was exactly a workshop on graph algorithms hosted at Rutgers University in the week between my two conferences. The program was almost exactly my research field, I was eager to hear from some of the speakers, and it gave me the oppurtunity to meet up with some of my collaborators that live in other countries. To get there, I took a direct flight to New York. After the workshop, I took the two hour train to Washington DC, where I got to spend the weekend with another friend. The next train, from Washington DC to Orlando had a special feel to it. Most Americans do not even realize this 20-hour train ride is an option, and book a flight right away. The train ride is also unfortunately much more expensive. Luckily, my university (the university of Salzburg) has the policy that you can always take a train, and if the journey is more than 8 hours, you can book a private cabin. (In fact, the university of Salzburg does not fund short-haul flights at all anymore.) This made my east coast trip a comfortable ride, where the private cabin is ideal for getting some work done. My train did not have Wi-Fi, and being disconnected can have the benefit of getting some productive hours in. In this case, I spend some time reading papers and working on my slides.  

Getting back home 

For my way back from Orlando, I could choose to fly to Salzburg via Frankfurt, or fly only to Frankfurt and take the train from there. For some capitalistic reason, the two flights were roughly 100 euros cheaper than taking only the first. Environmentally speaking, it is of course better to travel by train when you can. I chose this option, given the luxury that my supervisor’s grants can cover this and it is in line with my university’s travel policy. A big benefit is here that you can book a flexible train ticket; due to weather conditions the flight from Orlando to Frankfurt was somewhat delayed. That meant that some of my colleagues on the same flight missed their connection, delaying them by several hours. For me this was no issue, I walked straight out of the plane into the next train. 

CO2 comparison

To see what all this effort led to, let's do a rough comparison in emissions. The more conventional way of doing this journey would be flying Salzburg-Madrid, Madrid-New York, New York-Orlando, Orlando-Salzburg. I won't go into details here, but in the CO2 emissions are roughly: Salzburg-Madrid: 444, Madrid-New York: 1467, New York-Orlando: 419  Orlando-Frankfurt: 2363, and Frankfurt-Salzburg: 113 kg CO2. This adds up to 4806 kg CO2. If I replace the three unnecessary flights (total 976 kg) by train rides (total 150 kg), the total becomes 3979 kg CO2. Remember that we have 1500 kg CO2/person/year (see for example here). My personal conclusion from these sort of numbers is that you can travel as much by train as you want, while flying to another continent is something that can only be done every few years – definitely not a few times per year. 

Final reflections 

Although taking the train takes significantly more time, it is a more comfortable way of traveling – in my opinion. Moreover, the long travel times force you to rethink if the travel is actually necessary. Cutting unnecessary travel might be the biggest win for the climate in the end. 
On a bit of a side note, let me say that CO2 compensation schemes are not a valid alternative. Instead of me trying to explain this, I’ll refer to this Guardian article.
 
This journey has been undertaken within research projects receiving funding by the Austrian Science Fund (FWF) (P 32863-N) and by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No 947702).

IRIF lab now supporting conference publications without mandatory travel

Written by Antoine Amarilli
Classified in : Uncategorized Tags : none
IRIF  is a French research institute in theoretical computer science, whose members have been distinguished by numerous awards .

In a recent update to its internal regulations, announced publicly on its website, IRIF now openly supports the right of researchers to publish in conference proceedings while declining to physically attend the event onsite. Such decisions can be motivated by concerns on work/life balance or on accessibility, or can arise from the environmental impact of travel.

IRIF thus makes a public statement: being accepted for publication in conference proceedings should not require travel, even when conference organizers want to make onsite attendance mandatory. Of course, there are many solutions for conference organizers who want to follow IRIF's policy: allow prerecorded presentations, make the conference hybrid, make conference presentations optional, allow presentations to be given during another event, or another edition; or, of course, organize conferences without formal proceedings. It remains to be seen whether this incentive will initiate a change of practices in the community. Indeed, it would be hard for a major conference which acts as an important publication venue to openly reject such a policy: it would amount to declaring that, when evaluating submissions and researchers, the ability to travel counts more, or at least as much, than the quality of one's research. This would penalize women, young parents, colleagues from remote or less wealthy countries, as well as colleagues who do not wish to accelerate the climate crisis.

TCS4F, of course, supports this initiative by IRIF, and hopes that it can inspire other labs to adopt similar policies, and make it easier for researchers to publish without having to travel. TCS4F hopes that such initiatives can make conference publication, and ultimately careers in computer science, accessible to those who cannot travel or wish to protect the environment. 

For information, Thomas Colcombet, the current deputy director of IRIF, is also one of TCS4F's maintainers. Another TCS4F maintainer, Hugo Férée, is a lecturer at IRIF.

Taking action against the environmental crisis

Written by Thomas Colcombet
Classified in : Uncategorized Tags : none

As someone who is, as many colleagues, very concerned about the climate crisis, I decided to take some steps and change the way I consider my life as a (computer) scientist. This means that I care about my personal carbon footprint, and more globally the one of my community, and that I would like to have a positive impact on it. But, what can we do?

This post collects some ideas of actions that can be taken by academics if they feel concerned about the environmental crisis, and some related comments. It does not cover non-work-related personal choices that individuals can take, though there would be much to say about it. Nor is there any attempt to measure impact, or value the respective merits of each choice.

Let us emphasize that the choices made should of course be balanced with career considerations: the idea is not to jeopardize one’s future.

Personal changes

The first thing that can be done is to reduce your personal environmental impact as a scientist. To do so, you need to rationally estimate the impact of your actions, and find efficient ways to reduce the impact. This does not necessarily mean radical change, as there may be some easy gains.

Of course, in quantitative terms, your individual footprint is only a drop in an ocean if you consider your university or your community. But this step is crucial for several reasons:

  • It makes you feel better. The climate crisis is a huge and dire crisis, and the widespread inaction causes cognitive dissonance, so action is a necessity.
  • It makes you an inspiration to others, which leads to change at a larger scale. For this, extreme changes are not needed, and can even be counterproductive since examples that are beyond reach may be unappealing.
  • It can lead to discover new ideas and new ways of working.
  • It can help understand the difficulties that prevent others from changing.
  • It makes you think about the consequences of your actions, and makes you assess what is really important for your work.
  • Finally, and obviously, you cannot convince anyone to change if you do not walk the talk yourself.

But what concrete individual actions can you take? There are many common ways to act: reduce or eliminate meat or animal products in your diet (and require such options when participating to events), commute by bike or public transportation, reduce heating, avoid air conditioning, etc. However, as a TCS researcher, your main source of emissions is likely to be plane travel, so here are some ideas:

  • You can try fix yourself a maximal budget for plane travel, for instance: never, once every two years, once a year, or even twice a year while in PhD or postdoc. It is perfectly possible to avoid flights altogether, in particular if you are in a well-connected place with many conferences happening nearby (e.g., in Europe), or if you are at a later point in your career. It may be harder if you are very isolated geographically or thematically.
  • You can favor or even organize local collaborations and local events.
  • It is important to be aware of the right tools for remote collaboration: videoconferencing tools, virtual whiteboards, emails but also real-time chats, version control systems such as git or Subversion.
  • A possibility is also to prioritize conferences that are reachable by train or accept remote presentations. Publishing in journals can also be effective.
  • You may talk with your co-authors, and see if they would agree to follow the same policy for your joint papers. Clarifying this point ahead of time also helps. You may also decide to not invest energy in activities that would conflict with your beliefs, e.g., participating in the organization of conferences with mandatory in-person participation.
  • When you travel by plane, it is a good habit to combine scientific activities, e.g., conferences with research visits and other workshops. When you commit yourself to traveling less, quickly, maximizing the scientific reward of each trip becomes natural.

Communication and lobbying

While individual action is the most natural way to act, the biggest impact is probably obtained by influence other people to act. Here are some ideas in this direction:

  • An easy thing to implement is to display your commitment on your homepage. This should be done in a clearly visible way, so that people who visit your page for a specific reason (finding information about a paper, your email address, etc.) will see it. If you participate in climate-related initiatives or associations, mention it on your webpage and CV with links to give them visibility. You can also express commitment in research papers, e.g., as a footnote next to your name and affiliation.
  • It is also natural to mention your commitment in discussions. It is important to do so carefully, without cornering the other person into their contradictions, as this can be counterproductive: people cannot all invest the same effort, they are not at the same stage in discovering the problem, and they have different priorities. There can also be invisible reasons making action difficult for some, e.g., family commitments making shorter trips unavoidable, or psychological obstacles. The goal is to get everyone onboard, not to antagonize people.
  • An effective thing is to raise environmental related questions at conference business meetings and similar gatherings. This may seem intimidating, in particular if you are still a student: "Am I entitled to do this?" The answer is of course yes, and this is a compelling way to push decision-makers to consider these issues.
  • You can help conferences clarify their policy, e.g., by emailing organizers. For instance, ask before submitting your work if you can present it online. This is a legitimate question and is more visible than not submitting at all.
  • When giving talks, the use of examples related to environmental issues also conveys a message. This also applies to teaching, or when proposing project topics to students.
  • A more time-consuming activity is to produce material about climate change across various medias, such as writing blog entries, recording videos, etc.
  • Last, you can organize discussion groups in your university or department. Get involved in environmental action at all available levels (district, city, country, university, community, …). Many universities nowadays have a `green office’, and often also a scientists4future group.

The idea is to make your commitment visible, not necessarily to invest a lot of time in it. Again, this may look like a drop in the ocean, but at some point, when sufficiently many researchers will have expressed their will to change how research works, this will simply become the norm.

As a decision maker

As your career progresses, you get involved at many levels in the life and organization of research: steering committees, hiring committees, program committees, advisory boards of labs and research institutes, learned societies, etc. In these places, you may be able to change things at a large scale. But politics is tricky, initiatives often attract criticism, and decisions are met with resistance.

At this level, we can reflect about how science and research should take place in the future. What should be the role of conferences and workshops, formal or informal, compared to journals? How should we evaluate research, and by which metrics? What are the important topics and methods? The goal is to build a collective vision of a fruitful and fulfilling research activity, to which people can enthusiastically subscribe – not to advocate for painful constraints.

Here are some ideas:

  • In hiring committees, you can take environmental criteria into account. This may mean requiring environmental literacy in job openings. You can also ask applicants what is their vision and which action they are taking, and make this part of your hiring decisions. Beyond raising awareness, this topic is a very useful proxy to understand a candidate’s personal vision of research and judge their maturity.
  • When evaluating grant applications, you can value proposals that go in the right direction with respect to climate change.
  • In institutions, you can make it mandatory to account for CO2 emissions from travel. You can also try to change travel habits by:
    • communicating on internal channels and newsletters;
    • changing internal regulations, e.g., banning plane travel when a feasible alternative exists or for stays below a special duration;
    • adding financial overhead on plane travel, and allocating the surplus to climate-friendly activities, to achieve a redistributive effect;
    • choosing the right themes on which to prioritize job openings.

The problem is to have colleagues accept these changes. For this, it is important to ask people for their opinion, and to understand the diversity of situations that exist. On the other hand, equality is essential: nobody would accept to fly less if the person next door is exempted for an unclear reason. Often some political skill is needed to reach a common agreement.

As a conference organizer

Flying to conferences is an important part of the footprint of TCS researchers. Thus, conference organizers have a special responsibility in steering things in the right direction:

  • One first idea is to only help organizing conferences for which you are really enthusiastic about the topic and which have a clear role in the field. Indeed, because conferences also serve as publication venues, they can easily outlive their scientific usefulness. If a conference only survives because people have to publish, but would not attract any audience otherwise, then its existence can be questioned, and, for instance, its conversion to a journal-like format should be considered.
  • Another idea is to make it visible in the call for papers that the climate crisis is important and that the organization is adjusted accordingly.
  • In particular, you can enforce fully hybrid conferences. You can make it explicit that in-person travel is not required for speakers, and is not part of the decision of whether to accept or reject a paper. Make online attendance as compelling as possible. Indeed, mandatory travel to venues with formal proceedings makes travel unavoidable, in particular for students and early-career researchers. Online participation has other benefits, e.g., in terms of diversity or inclusion, for people who do not have funds to travel, have visa issues, family commitments, etc.
  • You can choose the location of conferences wisely: close to the center of gravity of its community, easy to reach by train and with other scientific opportunities reachable around (again by train). This needs to be balanced with the need to organize some events in more remote locations, but in all cases avoid touristic locations chosen for their exoticism.
  • In a similar way, you can reach out to relevant events (conferences, workshops, summer schools) in order to co-locate: with several related events occurring at the same place in consecutive weeks, participants can travel less for the same scientific benefit. Reach out to neighboring labs to promote schemes encouraging longer-term stays and visits, e.g., the Highlights and ICALP 2022 extended stay support schemes.
  • Last, you can experiment with new conference models: hosting the conference online every other year; bi-locating the conference (e.g., MFPS XXXVIII, with participants working during the morning in Cornell, and in the afternoon in Paris), running local hubs, extending journals with video presentations, journals with a fast selection process, etc.

See also our blog post on estimating the carbon footprint of your conference.

Be creative!

This is certainly the most important. There are solutions and things to be done that have not yet been discovered, and everyone can decide to take some percentage of their time to try them.

This blogpost is by Thomas Colcombet. It is co-signed by Antoine Amarilli, Hugo Férée, and Tijn de Vos.

How are TCS conferences adapting after COVID-19?

Written by Antoine Amarilli
Classified in : Uncategorized Tags : none

How are TCS conferences adapting after COVID-19?

Before 2019, the question of conference travel and its sustainability had been on the radar of many for several years, but with little concrete progress. Then, in 2020, the COVID-19 pandemic started: it brought international travel to a standstill, left people stuck at home during lockdowns, and academic conferences adapted on very short notice. A widespread solution was to move events online, replicating their on-site format as much as possible. The pandemic demonstrated that large-scale online conferences were in fact possible, though they raised many questions, e.g., about timezones or the feasibility of online socializing.

The COVID-19 pandemic is not over, but in many countries the situation is slowly coming back to normal, so that in-person conferences are again a possibility. We are now at a unique time for the academic conference ecosystem: Should we go back to conferences the way they were before the pandemic? Or should we adapt them with some of the lessons learned during that time? Arguably, we should aim for solutions that are scientifically effective (including networking and socialization), while addressing key questions such as environmental sustainability (e.g., plane travel) and inclusivity (e.g., visa requirements, lack of funding, caretaking obligations, etc.).

To understand the present trends, I survey here the current practices of conferences in theoretical computer science. There are too many conferences to investigate them all, so I made an initial list by picking those that currently having a representative in the recently-started open-access theoretical CS journal TheoretiCS. (This is just an arbitrary list, and not meant to imply endorsement by TheoretiCS or endorsement of these specific conferences.) I consider for each conference the upcoming edition at the time of writing, and/or the latest edition.

Our key criterion is the following: can you submit and formally publish at the conference in the usual way, but without traveling to the conference (e.g., with a remote video presentation, possibly prerecorded); and is this possibility made prominently clear before submission (e.g., in the call for papers). This was typically disallowed by pre-COVID conferences, which required on-site presentations except for last-minute emergencies beyond the control of the intended speaker. In my opinion, this criterion is a core issue, because formal publications are an essential requirement of career progression, especially for early-career researchers; and because academic conferences are the preferred publication venue in many communities, thanks to their prestige and predictable reviewing time. Thus, mandatory in-person presentation means that people who cannot travel or refuse to travel are excluded from the mainstream publication system. This criterion also echoes Moshe Vardi’s call in CACM for optional travel, shortly before lockdowns.

I do not examine the question of registration fees: it is fine for now if remote attendance costs the same as on-site attendance. Specifically, I agree that registration fees can exclude prospective participants, but the issue of how to bear the costs of conferences has always existed and I believe it is distinct. I am also interested in whether travel-free attendance is possible, i.e., are the talks streamed, again with or without registration: though again I believe that this issue to be less central than that of publication.

A disclaimer: this is an initial list, compiled at the time of this writing, and based on my best understanding. It may misrepresent the policies of some conferences, or the reason justifying some policies. It only gives information about individual events, which may be at odds with long-term plans for a conference, e.g., as decided by steering committees. More generally, at this early stage after the pandemic, I understand that many conferences are still experimenting and figuring out a new model, so the idea is to document, not necessarily to argue about what is right or wrong. If I did not accurately reflect the policy of some conferences, or if you have other feedback, I would be glad to know. My eventual hope in the longer run is that such a list can be kept up to date within the TCS4F initiative.

Purely online conferences

These conferences took place entirely online, so they allowed remote presentation and attendance:

Conferences allowing remote presentation

These conferences are fully hybrid, i.e., they allow remote presentation, and also allow remote attendance (even though they sometimes discourage it):

  • SoCG 2022, in Berlin and online, in June 2022.
  • ICALP 2022, in Paris and online, in July 2022.
  • COLT 2022, in London and online, in July 2022, which is hybrid though remote participation is discouraged: “The conference is hybrid in the minimal sense that we will enable authors to give their presentation-at-a-distance, and we will stream the conference and enable questions to be asked from a distance; but that is all. So on-line participation is possible, but support will be minimal: after over two years of pandemic, we want to very strongly encourage you to come in person if you can!” (source).
  • MFCS 2022, in Vienna and online, in August 2022.
  • CONCUR 2022, in Warsaw and online, in September 2022.
  • FSTTCS 2022, in Madras and online, in December 2022, but “In-person presentation is encouraged” (source).
  • ICDT 2023, in Ioannina and online, in March 2023.
  • FoSSaCS 2023 (part of ETAPS), in Paris and online, in April 2023. “We plan ETAPS 2023 as an on-site conference; nonetheless, remote attendance and presentation will be made possible” (source).

There are also conferences which will allow remote presentation, but for which it is not explicit yet whether they will allow remote attendance:

  • CSL 2023, in Warsaw and online, in February 2023.
  • STACS 2023, in Hamburg and online, in March 2023.

Conferences with no explicit policy

These conferences will feature remote attendance, but the call for papers and website do not seem to indicate explicitly whether remote presentation is possible:

  • ITCS 2022, in Berkeley and online, in February 2022. The policy for submissions to ITCS 2023 is also not explicit.
  • FoSSaCS 2022, in Munich, in April 2022.
  • PODC 2022, in Italy and online, in July 2022. The policy of PODC 2023 is also not explicit.
  • SODA 2023, in Italy and online, in January 2023.

Conferences with exceptional remote presentation

These conferences only allow remote presentation for situations outside of the control of the speaker:

  • STOC 2022, in Rome and online, in June 2022: “One author per paper is expected to attend the conference in person to present it unless there are strong reasons not to, including international travel restrictions due to COVID-19 or impossibility to travel for all the authors of the paper” (source).
  • FCSD 2022, in Haifa, in August 2022: “One author of an accepted paper is expected to present it at the (physical) conference, unless Covid restrictions prevent travel” (source).
  • TCC 2022, in Chicago, in November 2022: “TCC 2022 is an in-person conference. Speakers may be permitted to present live talks via Zoom only under extenuating circumstances, with permission of the program chairs” (source).
  • LICS 2023, in Boston, in June 2023: “Remote presentations can be organized for authors who are not to be able to attend the meeting” (source).
  • ICALP 2023, in Paderborn, in July 2023: “At least one author of each accepted paper is required to register for the conference, and all talks are in-person. In exceptional cases, there may be support for remotely presenting a talk.” (source).

Purely on-site conferences

These conferences have apparently resumed in a purely on-site fashion, with no online presentations or online attendance:

Acknowledgements

I am grateful to other TCS4F members, in particular Thomas Colcombet and Thomas Schwentick, for feedback and contributions to this post.

Rss feed of the articles