In this post I will just reproduce something written by someone else (with a bit of editing by me). I am posting it because otherwise it may not be posted at all and I think it may be of interest to some of us (where ‘us’ means researchers, especially in NLP/CL). So here is the first proxy post on this blog:
To Publish or Not to Publish
A friend of mine (let’s call him John Gourmet) recently had a major research paper rejected. The paper was central to his Master’s thesis and contained substantial innovation by all accounts (besides ‘impressive numbers’). John was so dejected that he mailed all his co-authors and supervisors that he wants to leave the whole publication business: forever. An excerpt:
I don’t bother about acceptance or rejection, but they completely ignored the novelty in the approach to solve the problem … I am really disappointed, sir. I am not interested in writing or participating in any publication work anymore …
That paper does not matter a lot to Gourmet as he will be defending his thesis soon (he has other publications) and is working in the industry. However, one of his co-authors replied:
… if you work more and come up with something more fundamentally innovative, your paper is likely to be rejected, and your own less good paper might be accepted … Don’t just stop thinking of publications. Without them you can’t survive in the research world. It’s like passing exams. Exams are stupid, you have to pass them if you want to do certain kind of work.
I would not discuss the importance of publications in the research world, nor the relevance of publications to a researcher. I have not seen enough and to some extent believe that publications are not exactly like exams. But the exam part of the latter statement really caught my attention. Which brings us to the reason why I am writing this.
Anybody who decides to go for graduate school should write something known as the General GRE. While it has some utility, I seriously wonder whether it significantly helps in separating the wheat from the chaff. I have known people who study for a single GRE test (which supposedly tests only basic verbal and analytical abilities) for as long as 3 years. How can one calibrate the performance of such a candidate with somebody who went through the testing pattern only a few times? It should be noted that the GRE is not supposed to be an exam which tests your preparedness or knowledge, but basic cognitive abilities. And students around the world (especially from India and China) regularly outscore their abilities (and their previous scores) by preparing long, very long.
Anyway, the reality is that we all have to face it: the standardized testing system. While many similar things in life can be hacked, standardized tests are far more susceptible to it. When I decided to go for graduate studies, I knew that I would be too engrossed in technical work and my research. Instead of studying for a longer period, I took out some time to figure a way out of studying longer. I prepared for around 3-4 weeks (only few hours daily) and got the scores I wanted.
Coming back to John Gourmet, one of the chief reasons why he is prone to not go for PhD is that he too is afraid of the standardized tests. Now he is afraid of the publication process too.
Just a reminder: This is a proxy post.