Special Interest Group on CRAP

For years I have listened to people argue that parallel sessions are harmful to scientific discourse, and how we need to maintain “quality”. While I strongly believe that quality of scientific publication should not be sacrificed, I think there is a harmful aspect to avoiding parallel sessions that is being overlooked.

The problem is that there are too many conferences. Just taking the example of the cryptography conferences and workshops organized by IACR, there were two conferences a year in the 80s, namely Crypto and Eurocrypt. Most people attended one or the other, and many people attended both. Then along came Asiacrypt, at which point most geographic regions of the world had a major conference once a year. The field continued to grow, resulting in more papers being written, and more people trying to work in the field. Rather than create longer conferences or have parallel sessions, the community created other conferences, including FSE, PKC, TCC, CHES, Financial Crypto, ACM CCS, Information Hiding Workshop, ECC, RSA Cryptographer’s track, etc. As a result, we ended up with much less of a sense of community, and subfields that often are out of touch with the mainstream of scientific thought on cryptography. This is not healthy.

This problem is not limited to cryptography – it extends to all of computer science. The root cause of this problem is our neglect for publishing in journals, preferring instead to publish our results in conferences. As a field grows, economic supply and demand dictates that we either expand the existing conferences or create new ones. The myopic view that parallel sessions will degrade any particular conference almost inevitably results in more conferences being created, and once a conference is created it almost never stops running (but sometimes dies a slow death of scientific irrelevance). There are no other fields that I know of who have adopted this model of conference publication, and it would be beneficial to the field of computer science research if we resisted this trend of creating more conferences.

By contrast, the field of mathematics has a single yearly conference in the USA that is attended by the majority of mathematicians in the USA. As a result, mathematicians maintain a strong sense of a broader research community, and cross fertilization between fields is facilitated. For those of you thinking about starting a new conference, think instead about how you can mold the conferences that already exist within your field, and start submitting to journals.