Here are a number of theory conferences that have changed their affiliation or are considering doing so in the near future. They are all somewhat larger than CCC in terms of attendance and budget, but still comparable. There are also non-theory conferences contemplating similar changes. One example is the much larger Conference on Computer Vision and Pattern Recognition (CVPR), which is considering leaving IEEE and becoming independent.
LICS
The Symposium on Logic in Computer Science (LICS) started in the same year as CCC, and was an IEEE-only sponsored conference until 2011. Since 2012 it is a joint ACM/IEEE conference. In 2012 ACM’s sponsorship was 25%, and this year it is 50%. The main motivation LICS people mention for the change is their desire to form a SIG for logic in computer science. They see the move as a first step towards that goal. Some members of the LICS community feel that their historic affiliation with IEEE was a mistake, but they cannot become a mere ACM conference. This is because ACM and IEEE have an agreement not to take over each other’s conferences. As a disadvantage of their move, the LICS organizers mention that the administrative overhead doubled due to having to deal with two organizations rather than one. LICS seems to care less about the financial and open access issues.
CoLT
The Conference on Learning Theory (CoLT) started in 1988 as a workshop, became an ACM conference (under the name “Conference on Computational Learning Theory”) in 1991, and decided to leave ACM and become an independent conference in 2000. They were unhappy about their interaction with ACM and felt that the service fee was unjustified.
They created a non-profit incorporation, the Association for Computational Learning (ACL), whose purpose is to facilitate the organization of CoLT. The ACL board plays a similar role as our steering committee, but ACL owns a permanent bank account and has a treasurer. I am still awaiting some more details, but ACL started with some donations and accumulated surpluses from CoLT registration fees over time. They keep about 1 to 1.5 times the yearly CoLT budget in their account. They do not buy any type of insurance.
CoLT has used a variety of venues for their proceedings. Since 2011 CoLT uses the Journal of Machine Learning Research, which is a free on-line journal hosted by MIT that allows authors to retain their copyright.
SoCG
The Symposium on Computational Geometry (SocG) started in 1985 and has always been sponsored by ACM, except for two years when it was organized independently but in cooperation with ACM. The first time this happened was in 2007, when the conference was held in South- Korea and the ACM requirements for sponsorship turned out to be incompatible with South- Korean law. The second time was in 2009, when the conference was held in Denmark and the organizers wanted to avoid the bureaucracy involved in ACM sponsorship. In 2011 the conference was held in France and the organizers wanted to have the same “in cooperation” arrangement, but ACM refused and insisted on the usual sponsorship arrangement. This was the reason for SoCG to consider terminating its affiliation with ACM.
Since then there has been a lot of discussion within the SoCG community, and negotiations with ACM are ongoing. The considerations of the SoCG community are very similar to ours. In particular, they care about open access. In addition, for the reasons mentioned above, they have issues with the ACM sponsorship arrangement when the conference is organized outside of North-America, which is every other year. Interestingly, they tried to negotiate an arrangement where the involvement of ACM is restricted to publishing the proceedings, but ACM declined. Another difference with CCC is that SoCG has a tradition of obtaining substantial gifts from local industry and other entities.
Next section: Options