Timing of CCC

In order to better serve the community, the Computational Complexity Foundation (CCF) is considering changing the timing of the Computational Complexity Conference (CCC). Since its inception, CCC has been held in the period from June till mid July, around the same time as STOC or somewhat later. Several years CCC was co-located with STOC, which facilitated attendance but precluded sequential submission to both conferences. Some years CCC was late enough in the season and STOC early enough so as to enable STOC rejects to be resubmitted to CCC, which helped in attracting submissions to CCC. For a couple of years, joint CCC/STOC submissions were allowed.

The timing of CCC was brought up a few years ago. The suggestion was to try and co-locate with STOC whenever CCC is in North-America, while maintaining the relaxed atmosphere of CCC. A new factor since then is the theory fest that has been developed around STOC. It makes co-location less desirable to some because the combined event becomes too long; some others would like CCC to become part of the theory fest irrespective. Joint CCC/STOC submissions were deemed undesirable for several reasons.

The goal of this post is to start a community-wide discussion about the timing of CCC. Here are three possible options that make sense logistically:

  1. Status quo (June-July)
  2. Late summer (second half of August end of July – August)
  3. Winter (January-February)

Here are some of the pros and cons of those options.

  • Status quo

    Pro:

    • Option of becoming part of theory fest.
    • Less conflicts with teaching and easier use of university facilities (lecture rooms, dorms).
    • Weather conducive to travel.

    Con:

    • Submission conflicts with STOC and ICALP.
    • Expensive travel.
  • Late summer

    Pro:

    • Avoids submission conflicts with STOC and ITCS.
    • Less conflicts with teaching and easier use of university facilities (lecture rooms, dorms).
    • Weather conducive to travel.

    Con:

    • Submission conflicts with FOCS, ICALP, Random.
    • Narrow Somewhat narrow window for scheduling the conference.
    • Expensive travel.
  • Winter

    Pro:

    • Avoids submission conflicts with STOC, FOCS, ICALP, Random.
    • Enables co-location with a complementary conference (SODA).
    • Cheaper travel.
    • Flexibility in review period (window of 4 months avoiding conflicts with STOC and FOCS).

    Con:

    • More conflicts with teaching and harder to use university facilities (lecture room, dorms).
    • Submission conflict with ITCS.
    • Winter weather.
    • Transition year with only half a year between successive CCCs.

We would like to hear what you view as the pros and cons of the various options. At a later stage we’ll have a poll to get a quantitative sense of the preferences within the community.

20 thoughts on “Timing of CCC

  1. I would vote for winter timing (though I am fine with summer too).

    I have another suggestion. Can we allow submitting papers that are currently under review in a journal?
    This might attract stronger submissions in those situation when the author(s) plan(s) to submit to a journal in any case, and would like to shorten the whole publication cycle.

  2. Dmitry, could you say a bit about the reasons for your preference? At this point, rather than votes, we’re interested in finding out what people see at the pros and cons of the various options. In particular, there may be some that we aren’t aware of or that we may have missed for another reason. At a later point, we’ll have a poll where we’ll collect individual votes.

    I’d rather keep this forum focused on the timing of CCC, but I think we have adequately addressed your other suggestion. Please check out the simultaneous submission policy in the CFP for CCC’18, and write me personally if you feel it isn’t adequate. I’ll bring it up with the CCF Board then, and we can discuss it at the CCC’18 business meeting if you want.

  3. Dieter, thank you!

    You are absolutely right, I simply missed the change of the formulation regarding simultaneous submissions (the last time I needed that “degree of freedom”, I couldn’t find any conference in the field that allowed it).

    Speaking of the dates, what I do find a bit inconvenient is the “floating” regime of the last few years. So, I think that making it more or less consistent is more important than the choice itself.

    The reasons why I personally prefer winter are subjecive: I agree with the above list of arguments, to me the “pros” of winter are the most convincing and the “contras” are less significant (the weather counts as an advantage – assuming it remains correlated with the seasons the way we are used to, of course).

    1. I fully support Dmitry’s point about consistency in timing. The timing of CCC has fluctuated in recent years, causing various different overlaps with other conferences’ submission periods.

      My other primary concern is to ensure there is the opportunity to flip either to or from at least one other major conference which is not too far away timing-wise, whether STOC, FOCS, ICALP, etc, as has been mentioned by various others in this thread.

      Finally, regarding winter time, for the quantum community, early-mid January is typically QIP time, just to add a data point.

  4. Is it obvious that late summer would cause conflicts with FOCS submissions? If CCC would be in, say, late August (and FOCS would stay in October), and if people on both sides are interested in coordinating, then maybe there could be a week or two between CCC notification and FOCS deadline? Or maybe not — I didn’t do any detailed calculations, but am just asking since the answer was not obvious to me.

    I personally kind of like the idea of co-locating with STOC, I have to say (and I was one of the persons voting for being part of FCRC, at least if I remember correctly, although I can also see the reasons against this). Perhaps I should make a second post after having attended CCC + STOC in a few months from now — I am looking forward to it and hope it will be great, but maybe afterwards I will just be totally exhausted and will never want to do it again. 😉

  5. The difficulty in fitting CCC after STOC and before FOCS lies in the review periods rather than the timing of the conferences themselves. This year there were 8 weeks between STOC notification (February 9) and FOCS submission (April 6). For a comfortable fit, we’d need 12 weeks: 1 week from STOC notification to CCC submission, 10 weeks for the usual CCC review period, and 1 week from CCC notification to FOCS submission. We could consider shrinking the CCC review period somewhat but not by enough, I think.

    Yes, I recall your argument in favor of FCRC :). Co-location with STOC certainly has its advantages, especially for people from overseas. That’s why two years ago we recommended to favor co-location with STOC whenever CCC is in the US (as long as we can maintain the CCC atmosphere). In the mean while, we got to realize the importance of one of the disadvantages, namely that STOC rejects cannot be submitted to CCC.

  6. A few years ago (around when I was “retiring” from the CCC conference committee) I pushed a bit on the idea of having CCC in the winter. I’m glad it’s being discussed! Here are some thoughts.

    – First, I would vigorously oppose the status quo and vote for any other option, unless there is some kind of forcing function to keep STOC’s submission/notification window and CCC’s window from overlapping (such as what happened this year). To me this is the strongest CON to the status quo, and it is very significant. I am certain that CCC suffers when this happens. In general, my support of any move of CCC is conditional on having a submission deadline *not* during Nov-Feb and *not* during Apr-June. SODA was very wise to deliberately not ‘compete’ with STOC/FOCS in this way, from its beginning.

    – A big reason why I thought winter is good: potential for co-location with ITCS and/or SODA. ITCS and SODA will be co-locating in 2019. Perhaps if it is a success (I have every reason to believe it will be) then that will add more evidence. Given the massive attendence I have witnessed at SODA over many years, it does not seem to be a huge barrier for theory folks to make it to a winter conference. I personally think a winter Theoryfest with SODA/ITCS/CCC would be a major attraction. All three are truly unique conferences at their core, and all three are quite different from STOC/FOCS in ways that are obvious to their attendees (even the ways they are different from STOC/FOCS are different from each other), so I think bringing them together would make for a very unique experience. CCC is a small cozy conference with a high-powered technical program, crucially having longer talks than any other conference I attend, giving more time in my mind to digest the technical stuff. SODA is a giant (in comparison) conference with a vast program, much shorter talks, but huge opportunities for outside discussion with many other folks. ITCS is small like CCC but broader than both SODA and CCC, with a very deliberate emphasis on new concepts and frameworks, with PC members introducing papers themselves, and novel side events. It’s three different ways to have a theory conference, and I wonder if each could learn some tricks from the other.

    – Another potential CON for winter that’s worth keeping in mind is CCC’s charter to be outside of North America on occasion (in which case co-location with SODA would be infeasible every year), necessarily making travel to CCC a long journey for some, every year. Winter could well complicate this.
    I sympathize, but also see some PRO in it. For one, I basically never have a chance to travel internationally in the winter, so I appreciate the novelty. More importantly, *because* CCC is often held far away from many attendees’ home countries, I expect travel to/from CCC would be *significantly* cheaper in mid-January, possibly leading to more attendees from opposing continents. (Personally for me, MIT has no classes during the month of January, so conferences in January are particularly nice for the faculty and students. However, for schools on the quarter system, January is much worse. In general, I don’t know how many universities are off for most of January, or not.)

    – Another PRO for winter which deserves more emphasis: a submission deadline during the summer. This makes it much easier to finally finish a manuscript that one’s been sitting on for a while (this is probably the #1 reason why I have a number of SODA papers…)

    – I was hoping the discussion might be also about early August, instead of late August. It seems to me that late August should include “POTENTIALLY MORE conflicts with teaching and easier use of university facilities (lecture rooms, dorms)” as a CON. Certainly this is true compared to the status quo. I think most universities (except for the few left on the quarter system?) start their academic year in late August or thereabouts, and I would imagine the disruption could only be worse than a disruption in January.

    – These various interferences with different academic calendars could be an argument for the status quo. At any rate, I’m for having CCC in the winter! If it were not possible to frequently co-locate with SODA and/or ITCS, then I’d lean towards keeping the status quo (conditioned on the first point I made, of course).

  7. I had a discussion with Paul Beame whether the review periods for STOC and FOCS may shift somewhat in the near future. It does not seem likely. This is both good and bad news for CCC.

    Good: The STOC notification date won’t move to later in the season (there was some rumor STOC might shrink its long period between notification and the conference). This means the above “late summer” option can be extended and run from the later part of July till the end of August without affecting any of the stated pros.

    Bad: There is no hope of fitting the CCC review period in between STOC notification and FOCS submission.

  8. I was originally opposed to winter, but I think Ryan convinced me :). (With all his caveats.) I think many universities on the semester system – even if they don’t have all of January off like MIT – still don’t start their Spring semester until mid-January. Which means having a conference in the first couple weeks of the calendar year often allows many academics to come. I suspect this is one of the factors that lets so many people attend ITCS, SODA, and, for that matter, the AMS Joint Math Meetings.

  9. Just to add very quickly that I have thought some more (not too much, though 😉 ) regarding the timing of CCC and actually am beginning to like the idea of winter more and more. Now, this might be partially due to that this summer is looking to be absolutely crazy in terms of travelling, but in general I find that summer is a challenging period to travel for family reasons. Sometime in January would be much easier in this regard. Turn of the month August-September would also be better than what we have now (but not earlier in August, which I believe is hallowed vacation season for much of Europe).

  10. I am strongly opposed to a winter timing, and agnostic about the late summer option.

    I feel CCC works pretty well right now. There might be some issues, but we should make sure to preserve what’s good about the conference when trying to address those issues.

    What might the issues be? Some people might feel that attendance should be higher. This is not a problem for me personally. I don’t have the attendance stats to hand, but the average number of submissions has been fairly stable over the years, and perhaps even gone up slightly. There is the sense of a conference in equilibrium. And indeed attending the conference reinforces this impression. It is relaxed, friendly, informal. I find it a much nicer experience than STOC or FOCS. In any case, there is space for many different kinds of conferences, and CCC is a good model for a specialized conference: large enough that it feels vibrant, small enough that people tend to know each other well and newcomers are made to feel welcome.

    Another issue is the conflict between CCC and STOC. I do think this is a more serious issue, even if the conference has worked well for many years despite it. The late summer option is one possible solution – a CCC/FOCS conflict seems less serious than a STOC/CCC conflict because of the smaller size of FOCS, and also because one would expect fewer CCC rejects to be submitted to FOCS than STOC rejects to CCC. But I also think we should try to explore more seriously the option to fit CCC between STOC and FOCS. The 12-week gap that Dieter says is necessary feels a bit conservative to me, but I do agree that 8 weeks between STOC notification and FOCS deadline is too little time. In any case, as mentioned above, the CCC/FOCS conflict doesn’t seem as serious. This gives more flexibility in scheduling the conference.

    There are several issues with the proposed winter slot. Some of them are already mentioned in the original post. But let me emphasize the issue with flexibility in scheduling. I think it would be a huge mistake to schedule CCC at a time that conflicts with teaching. But avoiding all conflicts with teaching would be difficult in winter as different institutions have very different schedules. We need to be conscious of the internationality of CCC – European researchers and researchers from Israel and India are very well-represented. This means that we need to take into account a range of different academic calendars.

    Speaking as someone working in Europe, travel presents an additional problem. CCC is held in North America twice every three years. Because there are fewer academic duties in summer, this still allows me to attend CCC fairly regularly. But if it moved to winter, I would find it almost impossible to attend North American CCCs. The winter weather would make travel an uninviting prospect anyway. And I suspect it would make it harder to organise CCC in Europe in winter. I am currently making a bid to organise CCC in Oxford in 2020. I definitely wouldn’t be making a bid if it were scheduled in winter.

    Finally, there would still be a conflict with ITCS and SODA, so we would be losing the relaxed summer atmosphere of CCC with very little gain.

    1. A few figures to give an idea of the fluctuations:
      – Attendance was 66 in 2014 vs 131 in 2016.
      – Submissions were 74 this year vs 110 in 2015.

      Also, the new timing (if at all) would go into effect in 2021. I did ask the three bids that I solicited for 2020 about the possibility of winter: one said no, and said yes, and one only wanted to do it in winter (not in summer).

  11. As some have asked to express my view – In my current assessment, the winter option is the best one. When we started discussing the timing a couple of years ago, I was reluctant to seriously consider the radical change that Ryan suggested. After pondering the pros and cons for quite a while, and having tried various local modifications, I became swayed that a substantial switch is justified.

    CCC vs STOC and FOCS:

    The main advantage to me is that complexity theorists will no longer need to select among CCC, STOC, and FOCS where to submit their work – just like algorithms researchers never needed to select among SODA, STOC, and FOCS. This is because the CCC review period comfortably fits in the period (FOCS notification, STOC submission) but not in the period (STOC notification, FOCS submission). This year the former period was 18 weeks, whereas the latter was 8 weeks. The usual CCC review period is 10 weeks, and to my mind we should leave at the least one week between notification and resubmission on both ends, resulting in a minimum period of 12 weeks.

    As for changes that may allow us to fit the CCC review period in the period (STOC notification, FOCS submission):

    – I asked Paul Beame about the possibility of making the period (STOC notification, FOCS submission) larger. It seems unrealistic to hope for gaining more than a few days.

    – One can consider shrinking the CCC review period. When everything goes smoothly, I believe 9 weeks is sufficient. Going much below that will come at a price in terms of quality of reviewing and is likely to result in problems some years. In fact, there have been years where the normal period was too short because of a serious hick-up, which caused a lot of stress for the PC chair and probably resulted in lower quality of reviewing. For that reason, we’ve actually considered increasing the review period to 11 weeks so as to build in a buffer.

    – One can consider shrinking the one week between notification and resubmission. I find one week already very little, and would not want to institutionalize the practice of resubmitting a rejected paper without revision.

    Having been closely involved with the organization of CCC for many years, I’m also worried about a tight timing that requires coordination between multiple conferences. We’ve had bad experiences with this in the past.

    Cost:

    A secondary advantage of winter is the lower cost of attending the conference. CCC makes an effort to keep registration fees low so as to enable researchers with little travel support to attend. However, the registration fee is often dwarfed by the cost of travel and accommodation, which are much higher in summer than in winter – a factor of 2 is not exceptional.

    Travel:

    Winter is less pleasant for travel than summer, but that has never prevented me from attending a conference. It has happened that I didn’t attend because of the cost of attending. I cherish the atmosphere at CCC, but I do not believe the weather is a significant factor.

    I agree that conference attendance is more difficult when teaching, but I’m not sure how many complexity theorists would go as far as not attending CCC because they’re teaching. That’d be good to know. For some, travel in summer may actually be difficult for other reasons (family, vacation).

    Overall, it doesn’t seem like any of those aspects have been issues for conferences like SODA.

    Facilities:

    Being held on a campus rather than in a hotel is one of the factors that contribute to the atmosphere at CCC. The availability of lecture rooms is likely worse in winter than in summer. However, many campuses have dedicated conference facilities, and those are likely easier to reserve in winter than in summer.

    Conclusion:

    The winter advantages of avoiding submission conflicts with STOC and FOCS, and cheaper travel are significant, and outweigh the disadvantages as I see them. The change is radical, but I believe it will bring CCC in a considerably better position.

  12. I’d also vote for any time other than the status quo, primarily because one has to choose between submitting to STOC or CCC. It would be great if CCC were held at such a time so as to have its submission deadline a week after STOC’s notification date, or a week after FOCS’s notification date.

  13. I think that the main factor that needs to be improved is allowing for STOC rejects to be submitted to CCC. I agree with Rahul that CCC rejects -> FOCS is secondary. Keeping the cozy and relaxed atmosphere is also very important.

    To me personally, winter travel is hard. Our quarter typically starts Jan 2nd, which means that travel would be in the first or second week of classes. When teaching large undergrad classes, it makes it essentially impossible. Also, it is much harder to use university conference rooms in the winter, as they are used for teaching. Finally, while summer travel is more expensive, as long as we are trying to run CCC at universities, we could try to use university student housing, which should be available in the summer.

    Regarding the status quo option – if we push the status quo from late June to mid/late July, I believe that it should allow for the STOC->CCC path. For example this year, this would have necessitated pushing CCC forward by 3 weeks, which would have placed it in mid-July instead of late June.

    So my personal choice would be summer (either option is fine) and then winter.

    1. Having attended 21 out of the past 22 CCCs, I much cherish the relaxed atmosphere of the conference, but I think the main contributing factors are the single track, longer talks, relatively light schedule, and campus location. Those wouldn’t change if we switched to winter.

      I agree that rooms that are used for lectures are harder to get in winter, at least on weekdays. However, several universities (including mine) have dedicated conference facilities, which can only be used for conferences and not for courses. Those are actually easier to obtain in winter.

      Each of the 6 years of my Presidency I have pushed local organizers about dorm rooms, but somehow it doesn’t seem that easy, at least not in the Status-quo – it only worked out once.

      I agree that pushing CCC back to late July would be enough to enable a proper STOC-to-CCC path. This is captured in the Late summer option.

  14. For me personally, summer travel is slightly more difficult than during the academic year because (a) my kids aren’t in school (possibly need to find childcare for them when I travel, depending on their schedule and my wife’s), (b) summer obligations that are not flexible (e.g., assisting with on-campus events for high school students, things with the family).

    Regarding having a conference during teaching time, I already travel roughly once per semester during teaching time anyway (mostly for regional events). I have even had to travel the entire first week of classes. I have been able to make it work with some front-loaded video lectures, a bit of email while away, etc. It is not ideal, but less of an issue than the difficulty of traveling in the summer.

  15. My preferred option would put CCC in competition with ICALP for submissions. This rules out June and STOC/FCRC co-location. Note that ICALP is often in early July so it doesn’t necessarily put things in “late” summer but I will vote for the “late summer” option to mean this.

    Weather is a major travel issue in Winter and the only way that SODA manages is that it only puts things in a warm climate. Many CCC conferences have been on university campuses in the northern US and Canada as well as in northern Europe. CCC in Riga was beautiful last year. Can you imagine that in January? We rely on academic institutions for locating our conferences and there are relatively few in warmer climes with a strong presence in Complexity.

    I also have classes starting Jan 2 and rarely have chosen to attend conferences like SODA/ITCS during this time of year because of it.

    If CCC does keep its current schedule, we should switch the submissions back to late November/early December so the committee work can be done over the Winter break and ICALP is an option for papers not accepted. The 2018 schedule conflicting with both STOC and ICALP/COLT etc was pessimal.

  16. In response to the poll among the CCF membership regarding the timing of future CCCs, 83 votes were cast. The counts for the 6 possible orderings of the three options (S: status quo, L: late summer, W: winter) were as follows (most preferable to least preferable, left
    to right):
    ■ SLW: 21
    ■ SWL: 12
    ■ LSW: 13
    ■ LWS: 10
    ■ WSL: 5
    ■ WLS: 22

    After discussion at the CCC’18 business meeting, the CCF Board concluded the following: “The Board recommends to set the CCC submission deadline after the STOC notification deadline for the next one or two years. The Board shall then evaluate the effects of the change and reconsider the timing of the conference.” See the proceedings of the CCC’18 business meeting on the CCF home page http://www.computationalcomplexity.org/foundation for more details.

Leave a Reply

Your email address will not be published. Required fields are marked *