How to write:
By Curtis R. Menyuk

A. Basic organization
1. Every scientific paper has an:
a. Abstract
b. Introduction
c. Main body
d. Conclusion
2. Every scientific paper has a point (or a theme if it is long).
This point or theme should be restated at least four times.
Bad papers do not beat the main point to death - this is not a mystery novel: subtlety is not rewarded!
3. All pages and equations should be numbered sequentially.
4. References:
a. Different journals/books use different approaches:
OSA: word# : # B. Gosh, R. Frame, and I. Whoever, "Trouble in OK Corral,", 13, 256-258 (1995), J. Opt. Soc. Am. B.
IEEE: word [#]. : # B. Gosh, R.Frame, and I. Whoever, "Trouble in OK Corral," IEEE J. Quantum Electron., vol. 13, pp. 250-254, 1995.

Some diversity exists for book references.
b. Be consistent!!
The best thing to do is to pick a journal or book and follow its format slavishly.
c. Pay attention to both in-text and final references:
IEEE:
(in text)
word [#].  (period follows)
word [a]-[b]. (not [a-b])
word [a], [b]. (not [a, b])
5. Figures:
a. For anything except a paper to a journal, figures should be embedded in the text on the page or page after it is first referred to. Every figure must be referred to. Later references in something long (like a thesis) can refer back to the figure's page number.
b. The caption's first sentence is a title and need not be complete. Following sentences must be complete (include subject and verb).
c. If the figure is taken from another source, it must be referenced like: (taken from ref. #) or (taken from [#].). If it is adapted, then taken -> adapted.
d. Figures must be clean and readable. If photocopying does not work, you have to redraw the figure!
B. Some common mistakes:
1. Overuse of acronyms:
a. Basic premise - it's a buyer's market and acronyms rarely benefit the reader.
b. Basic rule - never introduce a new acronym. If you must violate this rule, never introduce more than one or two in any text.
c. Differences between "networks" papers and "devices" papers:
(1) Acronyms are very widespread in the networks literature - at least give an acronym table.
(2) They are not widely used in the devices literature (with the exception of WDM or FWHM). Do not use them!!
2. Overuse of parentheses:
a. Parentheses interrupt the flow of the discussion and should be used very sparingly - always try to replace them.
b. Underuse of em-dash:
(1) There are three types of dashes
-     regular dash (the shortest) is used to hyphenate or join words
--     en-dash dash (longer) used to indicate number ranges e.g., 10 -- 20
---     em-dash dash (the longest) used for parenthetical statements
In Tex/LaTeX: regular dash = -, en-dash = --, em-dash = ---
In Ms Word: regular dash = -, en-dash = shift-, em-dash = Ctrl-
(2) In good writing, the em-dash is used more than parentheses because it is less disruptive of the flow (look at Asimov).
c. When parens are used, there is a space before and after but not inside (unless it is punctuated): word_(word)_word    or    word_(word),_word.
Never:    word(word),    word(acronym),    word_(_word_)
3. Overuse of italics and quotes:
- these should be used sparingly and consistently.
4. Overuse of i.e., e.g., and etc.:
- ditto for i.e. and e.g.
- Never use etc. unless you absolutely, positively can't think of a replacement. Remember, no mysteries!
5. Numbers in texts:
a. Below ten, they should be written out: There is one way to do it, two ways, and so on. Exception: reference to 1's and 0's in a digital system.
b. Large numbers (like 1,028) should be written out as numbers.
c. In between 0 and 100 is more iffy: be consistent. (10, 20, 30 - might be written out; 38 might be a number).
6. Improper referents:
__ Example: "Figure 2 is a network"
Figure 2 can "illustrate" a network or "show" a network, but it can't be a network. Think about what you are writing!
7. Hyphenation of adjectival expressions:
__ Used to indicate nouns that are used as adjectives:
"Mesa-type PIN diodes"
as opposed to: "These PIN diodes are the mesa type."
8. That vs. which:
__ "This is the one which I want."
should be
"This is the one that I want."
For native American speakers: If "that" works, use it to replace "which".
9. Misuse of spoken English:
__ People use networks "a lot"
should be
People often use networks.
Written and spoken English are not exactly the same. Particularly in scientific prose which is like a sonnet.
10. Overuse of we: (particularly in theses)
a. There are three principal uses:
(1) Referring to oneself: "We will show...". "I will show..." is better in a single person document.
(2) Referring to oneself and colleagues: "We showed experimentally...". That is right if you say earlier "My colleagues...".
(3) Referring to the scientific community.
"Substituting A into B, we find..."
should be
"Substituting A into B, one finds...".
b. This whole issue is controversial.
C. Mathematical expression
1. Variables should always be a single letter and should always be italicized.
__ In TeX: This is done by going into math mode; i.e., surrounding with $.
This is true for both displayed equations and text:
"V(X) = AT = 10,
where V is ..."
Numbers, parentheses, and words are not italicized. Neither are +, -, and =.
"V(X) = Voltage"
Voltage is not italicized. In TeX you must use \hbox or \rm.
2. Functions are not italicized.
__ "cos(X) = exp(-X2)"
Spacing is special, though. In TeX, it is best to use \cos and \exp which automatically gives the correct spacing.
3. Half spaces are used before and after = sign. Minus signs are en-dashes not ordinary dashes. In Ms Word, you must do this specially or use a good equation editor.
4. Units are not italicized; spaces should go before units.
__ "V(X) = 10 V/m, or = 5 µm" (in TeX, the "m" must go outside the $ $ sign).
5. Other special lettering:
a. Bold, non-italic is used for vectors.
b. Bold or non-bold sans-serif is often used for matrices.
6. Solidus is not used in texts, only in displayed expressions.
__ Instead of use "a/b".
6. Subscripts:
a. Words or parts of words are not italicized
nelectron    -    for electron density; the n is italicized, the electron is not.
nel    -    for the same thing and the same rule applies
b. Numbers are not italicized:
n1 or n2
c. Indices are always italicized:
Eijk = cos(idx/ky) cos(kdz/kz) cos(kdz/kz)
d. Single letter subscripts are italicized even when they stand for a word:
ne    -    for the electron density (I have seen some exceptions).
e. Multiple letters that begin two different words are usually italicized:
Vbi    -    for the built in voltage (caps for an acronyms would never be italicized).