As PhD students, our priorities always elevate our research.. Thus, our publications, as the proof of the achievements of our research, seem to be everything that we spend our time for. Considering such effort and time spent...
Why the hell that paper is rejected?!
One of the talks at Grace Hopper was, how to publish your research. The presenters were
Holly Rushmeier (
Yale University) and
Jaime Teevan (
Microsoft). It was a really enlightening talk regarding what we should and shouldn't do when writing, waiting and after the decision. I will summarize the talk, and also feed in my observations.
Before starting, I think the most important issue is the discussion of why we want to publish that much. I don't consider myself as a successful author, but I believe I have the correct motives (I believe more after the talk actually). The first or sole reason should not be "because I want more publication/reputation/citation/money/fame/...". It is because:
"The world needs your ideas!"
In that manner, the quality vs. quantity debate becomes more clear, as mentioned in the talk, it is good and advised that publishing as much as you can, but without sacrificing the quality. Three important and main points:
- Express,
- the great idea,
- well.
Depending on the area that are in, maybe you have lots and lots of good papers, i.e. in systems area. However maybe you work in theory and you proved three really important theories and tha your gift to humanity. The key is, none should spin around the same papers. And how you asses the success of a paper after it is published? The recognition you get is measured by the citation count. Even if you have small number of papers that are always cited from important people, then you're on the right track. Continuing with the thought flow, then how to ensure that they are actually cited? Submit to highly visible venues!
We have many options here.
- Conferences, journals... biggest outlets. I know that some of your friends might argue with you and underestimate the impact of conferences, but I guess they are not CS people. Let them shine in Science...
- Specialized workshops, accepting full papers or abstracts or posters. And guess what, you get to visit that small city in a state that you are not sure whether such a state exists! :)
- Software patents, data, code... Tangible things. If you can make your software fully available to the public, next to the paper you write, that is fantastic! So you both help people and get the recognition. Actually, as the perfect place, I want to thank all of the open-source projects and frameworks I have used, you always save the day.
- Social media? Blogs! Never thought of it in a formal environment right? Sometimes ideas need to be expressed in a timely manner, so social media is a great way to not wait forever... You write a blog, and wow I'm the 345th reader!
Conferences
As it was said before: CS is different. Historically; conferences become more and more important. Once upon a time, journal publications were really slow -I mean slower than now, or ever- and it could take years to see a really good idea published. So instead of journals, and their lazy reviewers, and the lazier editors, (nope, just KIDDING! I love you journal guys, that's why I keep returning to you... come on, don't be resentful), the conferences became more and more selective and getting more important. Biggest and hardest ones? For sure ACM sig's, i.e. SIGGRAPH... (As a note, I should apologize for my loud "YES!" from the people sitting next to me during the actual presentation. Sorry guys, I was excited.. to hear that it is hard..est...).
However, one should keep in mind that not every conference is great, there is a gazillion of them. How to differentiate is a whole new blog post.. but here are some questions that should be asked: do you need to just submit an abstract or full paper? how is the review process? how is the quality of the feedback? what is the number of reviewers? what it its acceptance rates for the previous years (probably this is the most representative question)? Under 50% is considered good? what is the number of citations does that conference get? and how often cited? what are its sponsoring organizations? Do they have ACM and IEEE? Or is it another yet important venue, like Eurographics (in graphics)....
Deadlines!!! For example for SIGGRAPH and Eurographics, they are really dead... The submission dates are absolutely strict! Not even one min after the deadline is acceptable. So prepare yourselves accordingly (is of course the advise, never the reality).
One of the best aspect of conferences: There is a certain date to get a decision, like your birthday:) It will come no matter what. You will face the truth on that day. However the truth is not always black or white, there is also a conditionally accepted decision, meaning "if only you do these, then it is yes!"
At that point, I see the phrase "SIGGRAPH deadlines:)))" in my notes, and I remember the timeline of SIGGRAPH being projected on the screen.. Watching it so many times during the process, I swear that I could repeat it my eyes closed... Anyway, thanks again for having all examples for SIGGRAPH, that motivated me personally x10 times!!
And we all know the process. Submit, reviewers are assigned, some of them you already know but none of them you will ever know, but good news they will know you if you get accepted, and they will never know you if you don't... And the waiting phase, more waiting. Also for some of the conferences, there is a rebuttal phase to answer the ridiculous things that the reviewers said, that you want to argue, but you cannot, that passes like a silent self-defense without being allowed to use any weapons, because yes you are not allowed to use anything but words. And every year it changes...
Coming to the ethics part, which we all know and the most boring to explain.. like reading the tutorial of the game that you have been playing all the time:) I will give you a skip option to next paragraph if you are already bored. SKIP. Rule 1: no dual submissions. Never and ever, send the same paper to more than one venue, because 1) it is asking for more review sources then you are normally allocated to. If the conferences say it is three reviewers, sending the paper to two conferences and having six reviewers is not fair, to nobody. 2) Every conference needs new and more developed content. If it is already published in wow-what-a-nice-conference-you-are then why also submit it to the-secondary-what-a-pity-conference?! If you are in doubt about whether it is a dual submission or not, you can always ask the editor/chair of a conference, it is better to act before than apologize and self-destruct later.
More ethics, if you get accepted to a conference, attend the conference, commit! Why not go?! People are there to hear you presenting it, because you did it, it's your precious work. And a better reason, conferences depend on attendances financially. If you cannot attend because of the government does not allow you to go (which is the only excuse from stopping you to go according to Rushmeier), as a last resort, find somebody that can present it as good and also attend...
Journals
Benefit: You don't have to go somewhere, yay!
But: there is no deadline.. The review cycle can be so slow, taking years...
How to decide that it is a good journal?
ISI Journal Impact Factor: nb of citations in year n to articles published in year n-1 and n-2, divided by total articles published in that year. (I believe my note-taking pace was not comparable to Prof. Rushmeier's speaking speed, so please don't believe me in this calculation). There is also H factors, Elsevier journals are good, and again you have to know your particular area.
The outcomes of journals are also a bit flexible, because of not having a publication/camera-ready deadline. As I said, there is a conditionally accept case for conferences, however for journals it is accept/reject/revision(major/minor). And guess what, since the authors are always expected to be capable of improving it, the accept never happens at the first time, perfectly. Minor revision means "probably accept it if you do these revisions", where major revision means "we didn't get your idea, but if you do these and these, maybe there is a cool thing behind your project?", and unfortunately reject means "No! Get out! Never ever come with the same paper again!", because revisions are so normal.
And the same review processes, probably double blind, which avoids bias and prejudices. However don't destroy your paper to make it totally totally anonymous. Thinking of which, who are those reviewers that we are so afraid of!! Yes, it might be you, me, the prof next door.. and also, the people you mention in the prev work gives an idea of whom to expect! (..if they are alive).
As many others, I was wondering how some reviewer minds are working, and here is their point of view. They look for your main contribution, what you are telling people that they do not already know, what solid evidence you have that shows it is valid. They also need to be given good reasons for both strengths and weaknesses of your work. So look for the good stuff. If it is a good paper, be explicit about that, people need positive comments, positive things about why it is important and good, don't be shy or humble, defend your years-old project!
Lastly, some more ethics: if you have conflict of interest with a paper, always turn it back, don't review. And yes, your best friend, or best colleague that you work on the same project, and your aunt are conflicts... And if you are reviewing it FOR someone, make sure that you get the credit you deserve about reviewing, on the list of reviewers for the conference.
Up to know, it is the first portion about the how/where/when/what to publish, based on the ghc13 talk of Holly Rushmeier and Jaime Teevan. In the second part, I will continue with the part-by-part analysis of an ideal paper. Stay tuned! I broke the post into two, because I know you have a paper deadline tonight and you need procrastination material... Right? :)
