Some guidelines for writing technical papers
E. Choueiri

Take the time to read the following guidelines regarding the writing
of a good technical paper. Following closely these general guidelines
will greatly reduce the effort and number of iterations needed to
shape up the paper and save you and me a lot of time.

Technical writing is not creative writing. I will never allow myself
to advise you on how to write a good poem or your memoirs, but there
are relatively clear and well-proven guidelines on how a good
technical paper should be structured and written. We are not talking
about great literature, but rather concise, clear and well-structured
papers.

I often find that students have problems with the abstract and the
introduction and consequently I focus below on these two topics.

The problem many students have with the introduction is its
structure. They often spend a lot words on some technical details and
do not even give an idea of how the paper will go about showing what
they intend to show.

The structure of an effective introduction section should be as follows.

I. Introduction
 
   I.a Motivation for work (from a very general perspective). By that
   I mean stepping far back and situating the importance of the
   problem within the general field of EP or whatever overall
   discipline the paper belongs to.

   I.b Review of previous work focusing on the remaining problems
   (questions or deficiencies) the present paper claims to contribute
   to solving.

   I.c. A statement of the paper's main question(s) and gaol(s),
   followed by a succinct description of the general method and
   approach to be described in the paper. No details on the results
   should be given except for possibly saying briefly, for instance,
   "We will show that...".  Remember results go in the abstract and
   not the introduction.

   I.d A brief section by section description of the structure of the
   paper.

The order of the sections should follow closely the spirit of the
approach stated in I.c. For instance if you state in the introduction
that you will develop a second-order perturbation method and apply it
to derive an expression for the saturation threshold, do not put the
section on the threshold derivation before that on the second-order
perturbation method.

The Abstract.

The abstract is the *only* part of the paper that will be read by most
people. Therefore it is extremely important. I often re-write my
abstracts a dozen of times.

Unlike the rest of the paper, the abstract should written in the
impersonal voice.  This derives from the fact that the abstract used
to be traditionally written by the publisher to catalog the work. The
rest of the paper , on the other hand, SHOULD NOT (see further below).

The first sentence of the abstract should be a very succinct statement
of the problem; almost a rewording of the title with a few explanatory
words (see example below, for instance). The following 2-3 sentences
should be about the adopted method and approach. The rest should be a
statement of all the *major* findings. An abstract without a specific
statements, albeit concise, of the findings will be flagged by the
editor of any good journal.

Unlike the Introduction, the abstract does not have to describe the
*general motivation* for the work (i.e. as in item I.a above). A
couple of brief sentences may be needed to state the motivation behind
some *particulars* of the technical work but you should avoid at all
cost implying that humans will one day reach Mars because you know
have a better understanding of how the current attaches on the
cathode.

Also, the abstract should not refer to future work unless it is an
integral part of the point the paper is trying to make (this is rarely
the case in a good paper since, by definition, a good paper is one
that tells a self-contained and well focused story).

If you cannot avoid using an abbreviation, it should be defined first.

Avoid like the plague naming institutions and programmatic details,
like for instance "the work was conducted at Stanford's high falutin
lab as part of a Phase XIX NASA-SBIR program".  This also applies to
the rest of the paper.

Avoid bibliographical references. At most one may be given when one is
really needed. (AIAA does not like even that).  (In the sample PoP
abstract below, a reference to a single experimental work was needed.)

Always write your abstract after you finish writing the first draft of
the paper. It is often that only then you have a real clue on what the
whole story is about.


Sample Abstract

A theory is presented of anomalous resistivity and particle heating in
current-driven plasma accelerators such as the magnetoplasmadynamic
thruster (MPDT). An electromagnetic dielectric tensor is used for a
current-carrying, collisional and finite-beta plasma and it is found
that an instability akin to the generalized lower hybrid drift
instability exists for electromagnetic modes i.e., with finite
polarization. Weak turbulence theory is then used to develop a
second-order description of the heating rates of particles by the
waves and the electron-wave momentum exchange rate that controls the
anomalous resistivity effect. It is found that the electron Hall
parameter strongly scales the level of anomalous dissipation for the
case of the MPDT plasma. This scaling has recently been confirmed
experimentally [Phys. Plasmas 5, 3581 1997]. Polynomial expressions of
the relevant transport coefficients cast solely in terms of
macroscopic parameters are also obtained for including microturbulence
effects in numerical plasma fluid models used for thruster flow
simulation.

IMPERSONAL VS PERSONAL
_____________________

Now on the use pronouns: the impersonal vs "we" in the body of the
paper.  Novice writers and younger grad students often think that the
use of the impersonal makes their work sound professional and
objective. There is nothing further from the truth. It often adds to
wordiness, awkwardness and imprecision.

For instance, the following sentence

"It was found that the temperature is a function of current."

may beg the question: who found this, you or the authors of the last
citation you mentioned?  It is more precise, concise and better
English all together to say:

"We found the temperature to be a function of current."

Remember, you are not writing a user's manual but reporting on your
research, (which has presumably consumed a good part of your life.)

The best written scientific papers I have read (see the all-time
classics on http://webserver.lemoyne.edu/faculty/giunta/paperabc.html)
do not shy away from using the pronoun "we" as opposed to the
impersonal "it" which often leads to awkward and less precise English.

For instance in Niels Bohr's classic 1913 paper "On the Constitution
of Atoms and Molecules" he starts his celebrated derivation by saying:

"Let us at first assume that there is no energy radiation."

instead of the more awkward:

"First, it is assumed that there is no energy radiation."

Today, as many scientists are not as well-read as they used to be (I
know some who shamelessly admit to have never read a novel), journal
editors often need to remind authors not to shy away from using "we"
and avoid using the impersonal voice as much as possible.  That does
not mean that you should not at all use the impersonal voice but
rather that you should write your paper in your voice and allow for
some exceptions when it is idiomatically justified (like in the last 5
words).