The Nastiness Problem in Computer Science

My personal favorite is the rejection letter that Ben Shneiderman received in 1972 from Communications of the ACM: "I feel that the best thing the authors could do is collect all copies of this technical report and burn them, before anybody reads them." http://www.cs.umd.edu/hcil/members/bshneiderman/nsd/rejection_letter.html

This is only a problem for academics. In the real world (industry), the customers stand in judgement.

This is an academic problem only faced by academics.

The more precious excerpts from refereeing that I've received today (about the paper at http://www.andrebarbosa.eti.br/P_different_RP_Proof_Eng.pdf):

"This is one of those proofs that is "not even wrong", quoting Pauli."

"I cannot follow the paper. It makes no sense."

I am a physicist but have entered CS and now publish in this field. I do notice the attitudes you describe and they scare me because I get the impression that every other computer scientist is very unsecure. Rude comments from reviewers are common and editors seem not to care. But more than so, it is common that reviewers are clueless and barely understand the paper they review. So, if one reviewer is rude and clueless, and two are knowledgable and positive, then the editor still mainly listens to the clueless one, simply because negative critique is more worth than positive in this field...

I have encountered the nastiness phenomenon in reviews. This can be frustrating, especially when malice is coupled with what seems a deliberate misunderstanding, i.e., a malicious interpretation of what is being said or done.

Case in point: In a manuscript we defined a precondition for set used in a certain method. Let us call this P(S). The existing method had a different precondition Q(S). We showed that Q(S) implies P(S), but not vice versa, and that P(S) is still sufficient for S to be useful. The review -- and I was assured by the editor they were an ironclad expert on the very topic -- was a blunt rejection saying (more or less) "the original condition is not broken, it should not be weakened".

It seems to me that malice is in the computer science community taken as a sign of expertice. Perhaps because in the past, some experts have been known to be malicious, some of us try to imitate them in the hope of building a reputation for themselves.

I wonder if our publication model, i.e. favoring conferences over journals, has something to do with this... Conference reviewing is inherently more adversarial and lacks the aspect of collaboration between the editor and author to create a worthwhile publication that seems important for journals (at least the ones I've dealt with).

I suppose it could also be other way around...Maybe we prefer the conference model because of our negativity?

As a reviewer and as an author, I get the feeling (in some cases I actually know) that some of my (co)reviewers did one of two things:

1 had someone else less/not qualified review the paper, without bothering to check the quality of the review,

2. reviewed the paper in the last possible minute, probably after several reminders from the Program Chair.

In either case, it is hard to get a fair review.

This is a great article - thank you for airing this issue. I recently sat through two talks by candidates for a professiorial post. I was frankly embarrassed by the agression in my colleagues' questions. The agressive questions came from exisiting professors, almost as if they were unwilling to have anyone join their exalted ranks.

I am from physics, it has its own share of nastiness, different from what you describe. Right now, I work in a research organization dominated by computer scientists and have hence, written and reviewed some computer science papers. At the risk of sounding haughty (I do not mean to), I would say:
As you've said, computer science is a relatively new field and physics far more mature. This not only means that computer science has more upstarts reviewing and writing papers, but that the quality of research varies from excellent to mediocre to rather poor. As opposed to physics or natural sciences, where almost all research in a field is of similar quality (with respect to maturity). Now you might say that is good or not good, I don't know.
Also, computer scientists have far more funds to publish and hold conferences (at exotic locations), leading in turn to lots more papers to write and review, and all the related rage. I wrote a paper once in two years and reviewed maybe a couple every year while I was in physics. I do some reviewing/writing activity every week in computer science research.
Any field which is still adolescent is bound to have its rage problems, so I do not consider it surprising that computer science research is the way you describe (some of it in conformity with what I've observed). Maybe there is too much money in it for it's own good. I don't know. Just saying.

One challenge this poses to program chairs is that we can be misled by nasty reviews vs genuine rejections, especially when the nasty guy works hard. A solution might be to publish reviewer stats, including how often a reviewer is the minority, average length of review, etc. Might improve behavior if we know that poor numbers could brand us out of prestigious committes. We all want to be on PCs, and then act as if we can't be bothered and are too busy! This needs to change. Coming from industry, I know that sticks work better than carrots!