Database Reference
In-Depth Information
I highly recommend having a sentence near the end of the introduction that
starts, “The contribution of this work is”, and of using bulleted lists if there are
multiple contributions. More subtle ways of stating contributions, using verbs
like 'present' and 'propose', can make it more dicult for readers and reviewers
to ferret out which of your many sentences is that all-important contributions
statement. Also, do not assume that the reader can glean your overall contri-
butions from a close reading of the arguments in your previous work section.
While it is critical to have a clear previous work section that states how you ad-
dress the limitations of the previous work, as I discuss below, your paper should
clearly communicate your contributions even if the reader has skipped the entire
previous work section.
I find that articulating the contributions requires very careful consideration
and is one of the hardest parts of writing up a paper. They are often quite
different than the original goals of the project, and often can only be determined
in retrospect. What can we do that wasn't possible before? How can we do
something better than before? What do we know that was unknown or unclear
before? The answers to these questions should guide all aspects of the paper,
from the high-level message to the choice of which details are worth discussing.
And yet, as an author I find that it's hard to pin these down at the beginning of
the writing process. This reason is one of the many to start writing early enough
that there is time to refine through multiple drafts. After writing a complete
draft, then reading through it critically, I can better refine the contributions
spin in a next pass.
IAmSoUnique:
Do not ignore previous work when writing up your paper.
You have to convince the reader that you have done something new, and the
only way to do that is to explain how it is different than what has already been
done. All research takes place in some kind of intellectual context, and your job
as author is to situate what you have done within a framework of that context.
A good previous work section is a mini-taxonomy of its own, where you decide
on meaningful categorization given your specific topic.
Proposing new names for old techniques or ideas may sneak your work past
some reviewers, but will infuriate those who know of that previous work. This
tactic will also make you lose credibility with knowledgeable readers. If you
cannot find anything related to what you have done, it's more likely that you're
looking in the wrong subfield than that your work is a breakthrough of such
magnitude that there is no context. Remember to discuss not only work that
has been done on similar problems to your own, but also work that uses similar
solutions to yours that occurs in different problem domains. This advice is even
more critical if you were lax about doing a literature review before you started
your project. If you find work similar to your own, you have a fighting chance
of carefully differentiating yours in the writeup, but if a reviewer is the one to
bring it to your attention, the paper is most likely dead.
Enumeration without Justification:
Simply citing the previous work is nec-
essary but not sucient. A description that “X did Y”, even if it includes detail,
Search WWH ::
Custom Search