A Brief History of SOCG and ACM

Before describing the issues at play in the current vote in detail, I’d like to sketch (my personal view of) the 30-year history of SOCG and ACM.  This history is long and complex, and some aspects are frustrating.

The Symposium on Computational Geometry was organized in 1985 by David Dobkin, Joseph O’Rourke, Franco Preparata and Godfried Toussaint.  According to the cover of the proceedings, this inaugural conference was “Sponsored by ACM SIGGRAPH, in cooperation with ACM SIGACT”.  Yes, SIGGRAPH came first; in the early years of the field, it was unclear to everyone what computational geometry was, exactly—a subfield of graphics? Of algorithms? Of numerical analysis? The connection to algorithms was not as solid as it seems now.  (Michael Shamos published his initial papers on computational geometry at STOC and FOCS, but only because those were the first conferences that his advisor David Dobkin thought of.) Interestingly, the proceedings were published as Proceedings of the Symposium on Computational Geometry. Apparently the organizers were not yet confident enough in the future of the conference to include the phrase “First Annual” in the title.

From 1986 to 2002, the conference was officially “Sponsored by the ACM Special Interest Groups for Graphics and Automata and Complexity Theory”, according to the proceedings covers. From 2003 through 2006, it was “sponsored by ACM SIGACT and SIGGRAPH”.  I don’t know why the order changed in 2003, but the official swap was already long overdue.

In 1993 and 1996, SOCG was organized along with many other computing conferences as part of the Federated Computing Research Conference (FCRC), which was sponsored by the Computing Research Association in 1993, and by ACM in 1996 (and later years). After two iterations, the SOCG community voted not to join FCRC again, primarily because the conference was significantly more expensive and offered insufficient benefit to offset that cost.

In 2000, SOCG was organized in Hong Kong by Siu-Wing Cheng and Otfried Cheong; this was the first SOCG organized outside North American and Europe.  At the time, ACM’s conference rules required organizers to open a local bank account in ACM’s name. When the organizers pointed out that following this requirement would be illegal in Hong Kong, ACM allowed them to organize the conference independently, without ACM supervision or support.  Nevertheless, the conference was officially sponsored by SIGGRAPH and SIGACT.

From 2001 through 2006, the conference was sponsored by SIGGRAPH and SIGACT, or vice versa.

SOCG was organized as part of FCRC again in 2003, after significant negotiations between the SOCG steering committee (led by Steve Fortune) and the FCRC organizers. The vote to rejoin FCRC was largely swayed by the promise of joint registration; anyone registered for any conference would be free to attend any other conference on that same day.  At the last minute, the FCRC organizers reneged on that promise; SOCG registrants were not allowed into other conference talks.  Moreover, SOCG was scheduled at the tail end of the multiday event, so that no other conferences ran in parallel with the second half of SOCG, leaving us stranded in an expensive conference hotel far from the city.  The community vowed never to rejoin FCRC again, ever, period.

Should I mention the official proceedings style that ACM foisted on us in 2001 that absolutely nobody likes?  I actually contacted ACM, in my official role as a member of the SOCG steering committee, and offered to help develop a LaTeX style that met the approval of both ACM and the research community. ACM’s response, which I reported at the 2003 business meeting, could be most charitably summarized as utterly dismissive. Would you like to download fixacm.sty? Just remember to delete (not just comment out) the line \usepackage[obey]{fixacm} from your .tex file before you send it to Sheridan.

At this point, there was already an undercurrent of frustration with ACM in the SOCG community, with a few radical voices half-jokingly calling for independence, but with no serious efforts to actually move away.

In 2007, SOCG was organized in Gyeongju, South Korea; the local organization team was again led by Otfried Cheong, who had since moved from HKUST to KAIST. Again, it proved impossible to implement ACM’s conference rules, not because it was illegal, but because all payments required ACM approval, and the normal protocol of submitting an invoice and being paid directly by ACM (11 time zones away) would have added an unreasonable delay to every purchase. However, ACM was unwilling to offer the same flexibility it had in 2000.  After discussions with ACM and the SOCG steering committee, the organizers proposed to organize the conference “in cooperation” with ACM.  In-cooperation status meant that the conference could be organized independently, without ACM supervision or support, but the proceedings could still be published by ACM.  To our surprise, it also meant that the proceedings were considerably more expensive, and that ACM would not include them in the regular shipments of ACM proceedings to libraries or to SIG members.  (I believe they relented on this final point when it was pointed out that as subscribers, they had already paid for the proceedings.)

SOCG was again sponsored by SIGACT and SIGGRAPH in 2008 in College Park.

In 2009, SOCG was organized in Aarhus, Denmark.  For many years the conference was organized on a three year cycle, with two years in North America followed by one in Europe, but by this point the frequencies had changed to every-other-year in North America.  Seeking to avoid unnecessary administrative and financial burden of ACM sponsorship, the local organizers once again applied for “in-cooperation” status, which was approved.

SOCG was again sponsored by SIGACT and SIGGRAPH in 2010 in Snowbird.

In 2011, SOCG was organized in Paris. Following the precedent set in 2007 and 2009 for non-American conferences, the local organisers applied for “in cooperation” status, and began organizing the conference under the assumption that this status would be approved. ACM refused.  The local organizers had already attracted financial support from local companies, under the condition that the conference was not otherwise financially supported; this money had to be returned when ACM insisted on sponsorship. Between the loss of local support and the required ACM overhead, the local organizers estimated that ACM sponsorship increased registration fees by €100.

At the SOCG 2011 business meeting, the local organizers first officially raised the possibility of moving SOCG away from ACM. An overwhelming majority of the attendees at the business meeting supported a motion for the steering committee to investigate becoming independent, drawing up a concrete proposal for organizing independently, and polling the larger computational geometry community via the well-established community mailing list. By this time LIPIcs had already indicated willingness to publish the SOCG proceedings.

An additional complication was that the sponsorship contract for SOCG 2011 was never actually signed by ACM.  INRIA sent the final version of the sponsorship agreement to ACM in late April; for reasons that are still unclear, ACM did not sign the agreement before the conference in early June.  After the conference, ACM refused to sign a retroactive agreement, which made it impossible for INRIA to either spend or return the balance of ACM funds.  An agreement was finally reached in June 2013 that allowed most the funds to be returned to ACM; the remainder had been reclaimed by the French government (as usual with accounts of French public agencies that lay dormant for more than 18 months).

In October 2011, the SOCG steering committee sent a detailed message to the computational geometry mailing list, outlining costs and benefits of ACM membership versus publication through LIPIcs and asking for votes on three options:

    1. I prefer to stay with ACM.

    2. If involvement of ACM can be restricted to publishing the proceedings, at low cost for SoCG, then I prefer to stay with ACM; otherwise I prefer to leave ACM.

    3. I prefer to leave ACM, and organize SoCG as an independent conference with proceedings published in LIPIcs and with financial backing provided through other means.

The final vote was 29 for option A, 48 for option B, and 47 for option C. The steering committee took the vote as a mandate to approach ACM investigating the possibility of a publisher-only relationship.

SOCG was again sponsored by SIGACT and SIGGRAPH in 2012 in Chapel Hill.  At the business meeting, the steering committee reported that although ACM was willing to offer some concessions, they were not willing to publish proceedings without formally sponsoring the conference.  They also reported that ACM affirmed that they were willing to agree, on occasion and with justification, to “in cooperation” status for future SOCG, if requested by the local organizers, but they would not guarantee that “in cooperation” status would be granted whenever SOCG organizers asked for it.  Finally, the steering committee announced that there would be a second and final vote on the community mailing list that fall.

In October 2012, the steering committee sent a detailed proposal to the community mailing list, again outlining costs and benefits, describing the results of their negotiations with ACM, and cautioning that a decision to leave has serious and complex consequences. The proposal asked for a vote for two options:

    1. I prefer to stay with ACM.

    2. I prefer to leave ACM, and organize SoCG as an independent conference with proceedings published in LIPIcs and with financial backing provided through other means.

The ballot also warned that if both alternatives received more than 45% of the votes, or if there were fewer than 80 votes, the steering committee would decide on its own. The final vote was 36 votes (42%) for option A and 50 votes (58%) for option B. The steering committee publicly announced their intention to move the conference away from ACM.

ACM’s response is best described as shock and outrage.

One major issue was that the second vote came almost immediately after a change in leadership at ACM and SIGACT, and the new leadership objected that they were not given sufficient time to respond to the issues raised in the voting announcement distributed by the steering committee. A second objection was that 87 votes, while higher than the limit imposed by the steering committee, could not possibly be representative of the entire computational geometry community. Despite these objections, the new steering committee (elected in March 2013) decided not to reconsider the vote unless ACM’s position on future “in cooperation” status changed.

SOCG 2013 was sponsored by SIGACT and SIGGRAPH in Rio de Janeiro.  At the 2013 business meeting, I announced that after long discussions with Paul Beame (the new SIGACT chair), ACM had agreed not to object to good faith requests for in-cooperation status when the conference is held outside the United States; in particular, SOCG 2014 is being organized in cooperation with SIGACT and SIGGRAPH.  Moreover, the SIGACT executive board voted to cover the additional cost of SOCG proceedings when the conference is held “in cooperation”.  Paul was also instrumental in reopening negotiations between ACM and INRIA to clear the SOCG 2011 budget.  Given the concerns about the previous vote and the concessions by ACM and SIGACT, I proposed a third and really-final vote at the business meeting, which received nearly unanimous approval.

And here we are.

Advertisements

About Jeff Erickson

I'm a professor of computer science at the University of Illinois, Urbana-Champaign and the chair of the steering committee for the International Symposium on Computational Geometry.
This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

4 Responses to A Brief History of SOCG and ACM

  1. Pat Morin says:

    If I understand things correctly, the main benefits provided by the ACM are that (a) they protect the organizers in case the conference loses money, and (b) the ACM brand lends some extra credibility to the conference, especially outside the community.

    The drawbacks are 1) increased bureaucracy and organizational overhead, 2) increased registration fees, 3) ACM keeps any profits, and 4) crappy-looking closed access proceedings.

    Any major points that I’ve missed?

    • Jeff Erickson says:

      There are a few other benefits:

      (c) For sponsored conferences, ACM provides advance financial backing, which can be used (for example) for reserving conference rooms and hotel-room blocks, and some liability insurance.

      (d) ACM will actually handle all expenses directly, simplifying the financial administration of the conference. (However, this is inconvenient-to-impossible outside the US, and it really only works if ALL expenses go through ACM. The workshops and YRF have so far been organized independently.)

      (e) ACM publishes proceedings in their Digital Library, providing citation links and (they claim) a more stable and usable platform than ArXiv or LIPIcs.

      (f) SIGACT provides some additional resources for student travel support and/or awards. (We used it for student travel this year.)

      (g) ACM may insist that we change the name of the conference if we leave. Whether they actually have the right to do this is debatable; what is NOT debatable is that ACM has lawyers and we do not. Changing the name of the conference may also impact its prestige, at least in the short run.

      Becoming an independent conference will require developing a reserve fund to offset financial resources ACM would normally provide. This likely means higher registration fees than with ACM for a few years.

      But I think the biggest issue is actually one of framing. Whose conference is it?

      ACM’s past behavior suggests they believe that SOCG is one of ACM’s many prestigious conferences, and that the research community acts on ACM’s behalf. From that perspective, uniform policies about conference administration, copyright, and even publication styles (however ugly we find them) make perfect sense.

      But I think most of the research community thinks of SOCG as the community’s conference, and that ACM acts on the community’s behalf primarily by publishing the proceedings. (The financial backing is helpful, but we’ve repeatedly demonstrated that we don’t really need it.) From that perspective, the administrative and financial constraints that ACM imposes are intensely frustrating.

      Neither of these one-sentence summaries is completely fair or accurate. Nevertheless, I think the conflict between these two perspectives is really what’s driving the community’s dissatisfaction with ACM.

  2. Vera Sacristán says:

    Pat says “(a) they protect the organizers in case the conference loses money”.

    I was co-organizer of SoCG 2002 in Barcelona, Spain. I would like to explain how ACM protected us that year.

    I managed to convince our university to open a specific account for the conference. Therefore, I was lucky that I did not have to do all payments through ACM. Nevertheless, I had to deal with ACM bureaucracy anyway, as they required a bunch of forms to be filled and approved by them.

    In particular, I had a long struggle with ACM before I got the conference budget approved. Everything was ok, but they insisted that we had to increase the registration fees. I kept answering and proving that there was no need. After a long negotiation, I was obliged to increase the fees, although I managed to substantially reduce the increase they asked for. This is how ACM “protected” the conference from loosing money in 2002.

    With this rise of the registration fees, the conference obtained more money than needed, and I had to send the leftover to ACM. It wasn’t much, but if I had not objected to their request it would have been much more. I did not like the way they managed this issue.

    Many years have passed since then. Things may have changed.

  3. Pingback: For Me, It’s About Open Access | Making SOCG

Comments are closed.