When I was a young researcher nobody taught me how to write a paper. Maybe, it is a sort of a test: if we prove we can write one, without any guidance, just taking into consideration other papers, we are ready to face the research world. In fact, many young researchers don’t even know what a paper is useful for. Their lack of understanding and absence of instructions bring them fear and low self-esteem to write one.
Recently, I received some papers to review and I could see a lot of evidences that novice researchers still make the same mistakes that I used to at the time of my debut. I have contributed in publishing 17 papers so far, and it gave me some experience to share with you about paper writing. My intention is to be clear and honest, so there is no rigour in this text and I don’t try to magnify the academic world with nice words.
A paper is a written report of what is going on during a scientific research. That’s why it is so boring to read. This report should be published somewhere. If we do not publish it nobody will actually know that we are active in research, thus we are not considered as researchers. If we publish, good! It means that other researchers have read and evaluated our text, identifying some innovation there. There are many kinds of publications. We can publish our work in conferences, workshops, journals, etc. Some of them are more accessible, others are more challenging. “Challenging” means that the researchers that will review our paper might be top specialists on the field, and if we are good too, they are our competitors. It usually happens when we submit to a journal. “Accessible” means most conferences and workshops. Renowned conferences work like journals, but a lot of conferences are not so rigorous and our papers might be reviewed by specialists, but not necessarily in our field, which will lead them to observe other aspects, like scientific methodology, text quality, structure and some logic in our words. The formula of a good conference is = tradition (several years) + strong reputation (stars in the program committee and hosted by renowned institutions) + blind reviewing process + strong sponsors (i.e. ACM, IEEE) + no submission extensions (important dates are like mountains: they don’t move).
In general, a paper has the same structure of how research is conducted. However, we don’t have to write this exactly sequence neither use them as our section’s titles. You can be original, but please, preserve the essence. This structure is as follows:
Abstract: Considering our incredible summarization power, we have to write a few lines about the paper’s contribution in order to convince the reader to consider the work. People who are looking for references for their state of the art definitively appreciate well written abstracts.
Introduction: We introduce the document explaining the context where the research is inserted, which problem of this context the work is going to explore, a summary of the solution and the consequences of solving the problem, concluding with a reading map of the rest of the document. The introduction does not give details about the research, but it should invite the reader to continue through the rest of the content.
State of the art: Depending on the maximum size of the document, the introduction can also contain the state of the art in case of small papers (up to 6 pages), but it is so boring to read that it’s better to put it in a separate section, giving the option to the reader to jump it. But only when we have more space to write (over 6 pages). The state of the art describes recent advances that have been developed elsewhere. These advances are cited to a) justify our decisions; b) show a gap that our work is going to fulfill, or c) show a similar work with a different approach to solve the same problem we are exploring. The content of our work is supposed to be new, built on top of the state of the art and the rest of the paper is to prove that. It’s important to mention that some authors prefer to write the state of the art after the contributions, but it is just a style of writing.
Contribution details: The most important part of the paper is where we describe the work we have done so far, and it should contain something new over the presented state of the art. This new “thing” should be described into details in order to allow reviewers to fairly evaluate our work, and give us a constructive feedback. The more distinguished is our contribution from the state of the art the more chances we have to get our paper accepted. That’s why it is called contribution. Contribution to the science.
Validation: Ok, we have done a nice contribution, described it into details already, but how can we prove that it is actually realistic and useful? So, we have to describe how we could validate our research, showing the procedures, the execution and the results. Validation might be the hardest part in a research because the contribution is just an idea and the validation is the implementation of this idea. The same way we cannot trust a software that was not tested, we cannot trust a research that was not validated.
Discussion of Results: The results obtained during the validation deserves additional attention. We can use them to compare with other papers’ results, analyze possible implications and so on.
Conclusion: Some people make a summary of the whole paper in the conclusion, but we have to summarize only our contributions, validation and obtained results. Introductory discussions and state of the art are irrelevant here. The conclusion also considers future works, which is actually what we didn’t have time to do, and probably will never have. Who knows? But at least, reviewers will not complain about those future works because they are out of scope.
Depending on the reader profile they will read our paper in different ways. Of course, we will be the first to read 100% of our paper, maybe many times, and the last one too. Nobody else would be so patient to do such a thing. The reviewers don’t have time to lose because they have 10 more papers to read, besides other several duties. So, after the abstract, they will jump to the conclusion, see the contributions, go through the results, and finally write the review. Conference’s attendees will read only the abstract. After being published, only young researchers and people out of the research community will read the paper’s introduction. For experienced researchers, our abstract should be enough because they know too much about the field and they don’t have time for bla, bla, bla. In summary, each part of our paper has a particular audience and we should consider them in the writing process. The only exception are journals and renowned conference reviewers, who will read 100% of the paper like we did, and that’s all.
To be honest, I don’t like to write papers, but sometimes we have to do what we don’t like to get what we want. During the PhD, publishing a certain number of papers is mandatory. The better is the journal or conference, the easier will be the PhD defense, because the jury is aware that they don’t have to be so rigorous if other researchers in the academic world already have. After getting the PhD, the number of references to our work becomes more relevant (other papers referencing our paper in the state of the art) than the number of publications we have. Seeing these as indicators, the number of publications shows our research activity and the number of references shows the relevance of our work.
Besides that, I also have problems to conciliate formal language, readability and deadlines. That’s why I love so much my blog, a place where I can express myself through my soul.