Leftover sausage

Since the SOCG conference banquet was scheduled immediately after the business meeting, there were a few issues that we didn't get a chance to discuss.

First, why was the conference so expensive? This was the most expensive SOCG in five years, including FCRC in 2003. (Coincidentally, or perhaps not, it was also the lowest attendance in several years.) Moreover, even FOCS 2004, which was held in a central location in Rome, included lunches and a banquet in the registration fees, and was organized by the same Italian company, had lower registration fees. Unfortunately, registration fees will continue to be outrageous—IMHO, anything over $100 is outrageous—unless the community commits to holding our conferences at universities (as in Brooklyn and Vancouver) using local volunteers. See John Iacono's comment on my earlier post.

There was a little bit of muttering about Sheridan's complaints about noncompliance with ACM's ugly LaTeX style, via fixacm. Apparently, ACM finally noticed. Fortunately, the complaints seemed to focus on the top half of the first page: title, authors, and addresses. The next version of fixacm (coming soon!) will pretend to conform a little better. (One paper in this year's proceedings imminentized the eschaton. Hoopla! Fnord!)

Marc van Kreveld wanted to raise the possibility of having a poster session in addition to regular papers. There are good arguments both for and against this idea. On the one hand, adding posters would increase the number of accepted results. According to Joe Mitchell, the PC could have accepted 50% more papers if there had been no schedule constraints. Increasing the acceptance rate would increase attendance, which would increase the visibility of the conference, which could bring even more people into the field. On the other hand, as has been pointed out at innumerable SODA business meetings (drink!), posters and short papers are seen as second-class publications. The implication is that the authors of posters and short papers woudl be seen as second-class citizens, not fit to polish the shoes of the authors of real papers. Unfortunately, this implication is fairly accurate. Also, increasing the acceptance rate would necessarily decrease the prestige of the conference in the eyes of hiring and promotions committees, who seem to prefer hard numbers even when they don't make sense.

Günter Rote provided a pointer to his paper statistics and comments from authors about the submission process.

Finally, and most importantly, there was no discussion of the theory community's efforts to increase NSF funding for theoretical computer science, as there was at the (also beer-free) STOC business meeting. One question in particular was never asked: Are we computational geometers still even part of the theory community? The answer should be a resounding NO!, followed by a slap to the back of the head—of course computational geometry is part of theory! Look, we have big-Oh notation! Unfortunately, reality seems to disagree. None of the new material on TheoryMatters mentions computational geometry at all, although it does mention another border community: machine learning. With few exceptions, the computational geometry community rarely submits results to STOC and FOCS; this was not true ten years ago. Lots of geometric algorithms are published at STOC/FOCS by people outside the SOCG community, but nobody calls them computational geometry. (Sanjeev Arora's TSP approximation algorithms are the most glaring example.) For many years, computational geometry has been funded by a different NSF program than the rest of theoretical computer science. (This worked to our advantage when graphics was getting lots of money, but that advantage is now gone.) At one infamous SODA program committee meeting a few years ago, one PC member remarked that nobody at SODA was interested in computational geometry*, they have their own conference, they should just send their results there. (This declaration led another PC member to resign.) Apparently, the divorce has been a complete success.

*Update (10 Jun 2005): Despite the opinion of one strangely deluded PC member, the SODA community has always been a strong supporter of computational geometry. At least one geometer has been on the program committee every year since its inception. Geometric algorithms have always been well-represented in the SODA program, sometimes more than any other topic in the call for papers. This is not true at STOC and FOCS.

Comments

we wuz robbed

While NSF funding is important, it only really matters to people who
are faculty at a US university. I do not think this is a suitable
topic for general discussion at a business meeting at an international
conference. However, perhaps we could arrange a special bitchfest at
the conference over breakfast where the relevant people could drink
beer and figure out why NSF does not want to give buckets of money to
a stagnant subfield where the major conference can only support 40
papers.

banquet was not included in the registration fees for FOCS - the price for it was EUR 75.00 and it was outrageous.

Even if I agree with some correlation between the registration fee and the number of attendence, I think what matters more is the location. Brooklyn was a winner because you got a huge number of researchers/students from the neighborhhod (NY, MA, NJ, PA, even the Washington DC area). Hong Kong, Pisa, or Korea will be loosers in that respect. For americans, the difference in the airfare is much more than the $100 difference in the registration fees.

[JeffE]

To the anonymous commenter who posted from an IP address at CNR in Pisa, where SOCG is currently being held:

"While NSF funding is important, it only really matters to people who are faculty at a US university." — I disagree. Funding agencies in other countries (and the European Union) are also rearranging their priorities away from fundamental research. The larger problem is increasing the visibility of our field among the general public. Attracting funding agencies are one aspect of this; attracting students to computer science (as opposed to Instant Millionaire Progamming™) is another.

"...buckets of money..." — Buckets? A few cups would be fine. Despite a flood of top-quality proposals, NSF hasn't been able to fund more than a a couple computational geometry grants per year since at least the mid-90s, and even those have almost vanished in the last few years. (I'm not including multi-displinary programs like ITR (which no longer exists in any case); I mean small single-investigator grants that pay for one PhD student for a couple of years, one month of summer salary, and maybe a new laptop.) This year, the only NSF grants for computational geometry will be in the CAREER program. We're not whining because we can't buy a third Mercedes; we're whining because we (or rather, our students) can't eat.

"...a stagnant subfield where the major conference can only support 40 papers." — Stagnant? What the hell are you talking about?

To be fair, the expensive registration fee is partly because of a weak Dollar vs. Euro.

--Stagnant? What the hell are you talking about?

Of course the field isn't stagnant. The point is that this is the message that a 40-paper major conference sends.

I've heard people joke about the "suicidal tendencies" of the computational geometry community. These tendencies include:

(a) small SoCG acceptance numbers. Students are discouraged by high rejection rates and people dabbing in the field are shunned away. [Contrary to what Jeff says this does not make the numbers any harder. The moment the word is out that SoCG is no longer a long shot you-better-know-someone-in-the-PC conference, the number of submissions would go up. ]

(b) no clear second/third tier conference (contrast this with e.g. complexity with CCC/STACS/MFCS or algorithms with SODA/ESA/(WADS/SWAT))

(c) plenty of weakly refereed conferences which are given very little weight by grant review panels such as the European Workshop, the Japan Workshop and CCCG.

(d) weak links to direct application areas such as computer graphics, GIS, robotics, computer vision, protein folding, computational chemistry, etc.

(e) low submission rates to FOCS/STOCS/SODA (as noted by Jeff).

No one is suggesting the standard be lowered. There were 65 papers worthy of publication, only 40 appeared. Is this a desirable situation for anyone?

--At one infamous SODA program committee meeting a few years ago....

Interestingly enough, computational geometry is explicitly listed in the SODA call for papers for at least the last ten years.

--The vote was 8 for FCRC, 57 for Korea, 3 for DC, and 38 for North Carolina.

Add that one to the suicidal tendencies of the CG community. Shun FCRC with its wider exposure for a far away destination? Yes, Korea has a strong and vibrant CG community, so SoCG would eventually had to be held there, but why on the same year as FCRC?

[Sanjeev Arora]

Jeff:

In the advocacy business, the only option for all of us is to get actively involved. See my comments in the discussion on your remarks that took place on Lance's complexity blog.

Young and ambitious researchers such as yourself (and other readers of this blog) should take a lead in crafting a compelling, forward-looking vision. Feel free to join the TCS advocacy too.

[Vladlen Koltun]

"One paper in this year's proceedings imminentized the eschaton. Hoopla! Fnord!"

Kudos on the Robert Anton Wilson reference. BTW, which paper do you mean?

[JeffE]

Hey, I'll never... never... aaaahhSHEWchuk. Excuse me. What was I saying? Oh yeah, I'll never tell!