Writing a Paper - Writing for Computer Science - page 55

the problem is interesting, what the relevant scientific issues are, why the approach
taken is a good one, and why the outcomes are significant.
That is, the introduction should show that the paper is worth reading and it should
allow the reader to understand your perspective, so that the reader and you can
proceed on a basis of common understanding.
The introduction can discuss the importance or ramifications of the conclusions
but should include only a brief summary of the supporting evidence, which the
interested reader can find in the body of the paper. Relevant literature can be cited
in the introduction, but specialized terminology, complex mathematics, and in-depth
discussion of the literature belong elsewhere.
A paper isn't a story in which results are kept secret until a surprise ending. The
introduction should clearly tell the reader what in the paper is new and what the
outcomes are. There may still be a little suspense: revealing what the results are does
not necessarily reveal how they were achieved. If, however, the existence of results
is concealed until later on, the reader might assume there are no results and discard
the paper as worthless. 2
By the end of the introduction, the reader should understand the scope of the
work, and of the problem. For example, if the topic is “mechanisms for collaborative
authoring”, then you need to explain who is doing the authoring; what abilities and
experience are assumed to have; what kinds of tasks they are trying to complete; and
how sophisticated the mechanisms need to be.
The reader should also understand the contribution, that is, what the discovery
of the work is—the core idea that the referees or examiners need to appreciate as
novel and important. This understanding requires that the reader appreciates what
the properties of this contribution are, what makes it interesting and plausible, what
method was used to investigate it, and why the method is appropriate.
Body
The body of a paper should present the results. This presentation should provide
necessary background and terminology, explain the chain of reasoning that leads to
the conclusions, provide the details of central proofs, summarize any experimental
outcomes, and state in detail the conclusions outlined in the introduction. Descrip-
tions of experiments should permit reproduction and verification, as discussed in
Chap. 14 . There should be careful definitions of the hypothesis and major concepts,
even those that were described informally in the introduction. The structure should
be evident in the section headings. Since the body can be long, narrative flow and a
clear logical structure are essential.
2 There is an irritating kind of paper in which the reader's interest is baited with comments such
as “the analysis, as we show, led to surprising insights” or “as discussed later, this decision had
unanticipated benefits”, with no hint as towhat the surprises or benefits were. This is not an endearing
style of presentation.