The introduction of a paper is where we present the issue we deal with, we put it into the proper context and we state in what sense it will contribute to the scientific field, or even to society or to humanity. In other words, the introduction is the place where we say what we do and why it makes sense. There are different ways of starting the introduction: let us see some examples taken from a communication exercise I did with my students. I gave them a scientific paper (on data science) and asked them to re-write the introduction in different ways. We can start with a quite reasonable example. “Today a huge amount of data is open to the public (the so called Open Data). This data may come from different sources and have different structures, so data integration is necessary. The aim of this paper is to highlight a new paradigm for thinking about data integration.” In this example, the field of interest is introduced (open data) and an issue is highlighted (the heterogeneity of the data), leading quite naturally to the paper’s aim: finding a solution to the issue. But something is missing. Try to think about it. What is missing is why it would be relevant to enhance our way of dealing with data, the “broad” vision. You may say: “It can be taken for granted”. Well, of course. So or… maybe not? The answer would be “yes” or “no” according to the venue, the intended audience, the common ground, etc. What matters is to be aware of the fact that the broad vision is missing and that you may want to add it. So let us see another option, where a glimpse on the bigger picture is offered. “The goal of this work is to provide the principles and systems to allow data scientists to effectively find and integrate open data and make private repositories more open […] Open data is chosen so that we can share our empirical data and results with the scientific community, companies and practitioners to advance research.” In this example, an attempt at explaining why the work presented in the paper makes sense is made. Let us move to a completely different strategy: a “descriptive”, almost educational, approach. “In data science, one of the main challenges is Data Integration, which is a way to create a database adding together data from different databases. The integration is driven by data analysis, which requires algorithms that search over repositories containing huge tables. Data can be extracted from ‘data lakes’, which are repositories with a massive number of different datasets. Among data lakes we find: open data, i.e. sources of data that must be always available; mass collaboration projects, i.e. tables built by a huge number of users; enterprises data warehouses, i.e. corporate databases. There are two problems in integrating data from data lakes: finding joinable tables and finding unionable tables. This paper proposes a new way to find joinable tables…”. As you can see, a long explanation of the topic (please note: of the topic, not of the problem) is offered and only after a long number of lines the paper’s core issue is introduced. And now there comes the last examples. Please note that they still make reference to the same paper! “What were and are the main problems of Data Science? In the 80s, the main concern was to centralize data; in 2000, the main issue was to map and exchange information between autonomous systems. Nowadays, the amount of open data is so large that the main problem is how to retrieve the useful ones…” This version assumes that the reader’s interest is on the Data Science problems: the rhetorical question in the first line does not ask “Are you interested in Data Science problems?” but “What are and were Data Science problems”. Then, it leads the reader step by step, in a kind of historical perspective, towards the current problem. Assuming, as I said, the reader’s interest, the communication goal seems to be to show how the Data Science problems have somehow paradoxically evolved from lack of data to too much data. Eventually, the last example is: “The problem of data integration is changed with respect to the 1990s…”. This example is similar with respect to the previous one in that it assumes an historical perspective and – even more than the previous one – it assumes an interest by the reader in the data integration problem. We can now draw some lessons. The main point is: do we want, do we need to raise the reader’s interest or not? Descriptive/educational approaches take it for granted that the reader is interested and that all that needs to be done is just describe the work. This is the typical approach by students, for example. What we may call “problematic” approaches take the burden of raising the reader’s interest and showing why the work is beneficial. Problematic approaches can take many forms, according to what problem they identify. The simplest way is to identify a gap in the literature. Something like: "Topic X has been widely studied, but its application to Y is still, quite surprisingly, in its infancy!" In this case, you may also want to add why the application to Y would be so interesting. Make the reader feel we cannot live without applying X to Y. If you want to take a broader perspective, you can start from a social issue. Something like: “Society is suffering from the problem X; the research presented in this paper addresses the problem X in a super effective and efficient way…”. Problematic approaches are meant to “gain the reader to the cause”. Another way of gaining the reader’s interest is by showing that your work is a good opportunity. This can be done, for example, via an analogy or by identifying a paradox we solve. For example: “Human beings have got a number of senses, but traditional learning activities rely on two only: sight and hearing. The approach presented in this paper shows how robotics can introduce a more comprehensive approach to educational activities, in which more senses are involved…”. The last option, untypical for scientific writing, is what we may call the “situational incipit”, where you start with a short story or an anecdote, thanks to which the reader slowly discovers the issue dealt with and – most important – gets interested and even affectionate in it. This style is common in magazines, and it works really fine. I personally think it could be also used in scientific writing. Imagine for example starting a paper on a project for schools by telling, in a few lines, the “story” of how Gabriel, one of the students, experienced it. Stats, graphs, etc, will come later to corroborate the paper’s scientific credibility, but the story would surely capture the reader’s attention.