|Posted March 13, 2014; updated March 18, 2014 © Paul Beame|
I am taking no official position on the vote itself.
Since ACM and IEEE have many joint conferences, they don’t try to steal conferences from each other. As a member of the SIGACT leadership, my taking an official position is probably not appropriate.
What I think overall.
As someone who has been involved in the leadership of both ACM SIGACT and the IEEE TCMF, as well as being involved on the local arrangements side of 3 FOCS conferences and one STOC conference, I see no downsides and some potential advantages for CCC versus the status quo if CCC becomes jointly sponsored with ACM.
Current status of the relationship between CCC and ACM SIGACT.
CCC is currently described as in-cooperation with ACM SIGACT. This lets CCC mention this fact in its advertisement and materials. Included in this relationship is the fact that ACM/SIGACT and IEEE members get reciprocal member registration rates at FOCS/CCC versus STOC. Formally, this has no other meaning since CCC is currently 100% sponsored by IEEE.
What backing would ACM SIGACT provide?
In addition to a long history of successful conferences, SIGACT has been receiving a substantial sum of money each year from ACM as its share of the digital library revenues based on the downloads of SIGACT sponsored materials. While this amount could drop to 0 overnight if ACM policies on open access change, this money sponsors quite a lot of what we do in awards, student travel support, and the like. Moreover, even if it goes to 0, we have a large amount of money in the bank. This year, the DL revenues, together with some unusual conference surpluses, have meant that there is direct support for student travel grants and awards to SIGACT sponsored conferences in proportion
to their size. For example, LICS 2014, as a co-sponsored conference, will receive $4.5K for that support which won’t need to be in their conference budget.
How would a jointly sponsored ACM/IEEE conference actually work?
- There is no double paperwork
Each year, either ACM or IEEE would be the administrative lead. All paperwork would only be done with one of the two organizations, depending on the year. The organization in the lead then simply informs the other. There should be no extra overhead. (There was a communication error for 2013 LICS which is the basis for the erroneous general assumption in the manifesto.)
- How that lead is chosen
In order for the administrative work involved to be balanced it typically alternates between the two. ACM and IEEE have somewhat different models of how things work which makes it a preferable to have ACM as lead for conferences in North America. This can be arranged conveniently to match the FCRC conference cycle. ACM has an arrangement with regonline (our usual registration service) that halves the costs and sends the money direct to an ACM sub-account for the conference. Expenses are paid by ACM staff as authorized by the organizers. This means that organizers do not set up bank accounts or handle money and makes reporting easier. Though this can be done in a wide variety of currencies it is not so convenient outside of North America, when the IEEE model seems more convenient. Though there is some flexibility in alternation it seems that it is not possible to settle on an unbalanced choice of leads.
- How joint affiliation would affect budgets.
To support the central organization, ACM includes an allocation fee of 16% of all expenditures of its SIGs (above a certain level the rate starts to decline). Conference budgets are part of this so the required allocation is passed on by SIGACT to the expenditures of the conferences it sponsors. This is formally independent of the services that ACM provides. IEEE has an overhead of 20% with a minimum of $5K per conference that is viewed as being associated with services provided.
Each year, independent of who is the lead, the overheads and surpluses would be allocated 50-50 between ACM and IEEE. The overhead would mean 8% for ACM and 10% (min $2.5K) for IEEE. 100% of surpluses on the ACM side would go direct to SIGACT. SIGACT has arrangements with some of its conferences for 50% of these surpluses to be returned directly to those conferences and this seems reasonable should CCC become jointly sponsored. On the IEEE side, if the conference is closed in a timely manner (6 months), then 50% of the surplus becomes available to the TCMF for its use two years later, though this has to be budgeted and spent on short notice; otherwise no money is returned.
- Contingency in conference budgeting.
Any difference in these figures between IEEE and ACM is a red herring. IEEE requires a 10% contingency and ACM requires a 15% contingency in conference budgeting. However, none of this money is actually taken out in either case. Even without changing the number of attendees, it is easy to meet these numbers by adjusting the assumed distribution of registrants to include a higher percentage of non-members or late registrants, or a smaller percentage of students to compensate for much of the contingency. In the case of ACM, it is also possible to approve budgets with lower contingency percentages for conferences that consistently have budgeted well. Contingency is sometimes necessary. We did have one case with FOCS 2010 where even with a 10% contingency we would have had a small deficit except that SIGACT donated a few hundred dollars to make the conference end in the black given that consistently ending in the black is an important consideration the evaluation of conferences at IEEE.
Cutting things close works fine if you have an organization with deeper pockets backing you up. An independent conference could not cut it so close with the contingency and would need to build in a real contingency that is quite a bit larger than what actually gets budgets for either IEEE or ACM.
- Proceedings would appear in both organizations’ digital libraries.
This is already happening for LICS. Conference proceedings since LICS 2012 appear in both the ACM DL and CSDL/IEEExplore. The production of proceedings would follow ACM standards in years when ACM is the lead and IEEE standards when IEEE is in the lead.
- Proceedings costs and open access are separate issues
Though IEEE by default recommends IEEE CPS for proceedings preparation and ACM recommends Sheridan Printing, neither of these is actually required and they are considered vendors somewhat like the vendors for conference A/V or other optional services. What ACM and IEEE require are specific formats for the materials, table of contents, front matter, index, etc. They also require specific copyright transfer (or exclusive license in the case of ACM) as well as the transmission of either electronic or hard copy for inclusion in their respective digital libraries. In particular, none of the proceedings cost relates to the direct costs of either IEEE or ACM including the proceedings in the digital library; those costs are absorbed by the respective libraries. (I did not realize this until recently.)
- ACM and IEEE have similar perspectives on informal conference websites.
These websites include freely available copies of the papers, videos, etc. You can check out http://ieee-focs.org for an IEEE example, or http://sigcomm.org/events/sigcomm-conference and click on the program for one of the past conferences to see an ACM example.
What might the relationship be between ACM SIGACT and an independent complexity conference?
- The current “in cooperation” agreement with CCC would not automatically carry over to an independent complexity conference. Such an agreement would have to be renegotiated but I don’t see an inherent problem with this.
- In cooperation conferences like SODA can choose to be included in the ACM digital library. With such an agreement it would be possible for a new complexity conference (in whatever name it is able to use) to be included even if it also does open access publishing provided that the materials are produced to a quality standard.
- Other than this, the conference would be on its own and have to maintain its own long term structures, budgets, and volunteer organization.
Correcting some inaccurate or misleading information in the Manifesto and the update.
- There is no double paperwork for joint conferences.
The problem that LICS encountered was an error that has been corrected.
- The SoCG website to which the Manifesto and update refer as the most current does not include the latest information as of SoCG 2013.
The future affiliation of SoCG is open and there is a plan for a re-vote. Many of the concerns that precipitated the original vote have been addressed and the relationship with ACM-SIGACT has improved substantially, though I do not intend to pre-suppose what the outcome will be.
- Proceedings costs are not IEEE costs (updated).
In an earlier version of this response, I incorrectly stated that they were included in the ‘paid to IEEE’ number, but they are not. I am sorry that I misunderstood the spreadsheet; the fact that they were included in the table along with the actual IEEE costs but separate from other costs, which are not listed at all, is what gave me that incorrect impression. Prior to 2011, the costs in the table this included production of print and other hard copies of the proceedings. It is important to note that neither IEEE nor ACM requires the use of their publishing service to produce those proceedings. They only require the copyright transfer (or exclusive license in ACM’s case) and production to standard formats, however produced. These costs are an independent question.
- The contingency differences between ACM and IEEE are irrelevant.
- There is no difference between IEEE and ACM about informal conference websites that include papers in their programs.