EECS 556 writing guidelines

• Everything you submit should have a title, author names, date, and subtitle indicating its purpose, namely "project proposal" or "progress report" or "final report."
• If you cut-and-paste equations or figures from other sources then absolutely you must cite those sources every time. Use the LaTeX \url or \href commands in the hyperref package.
• Make a high quality bibliography where each reference has the complete list of authors, title, the complete citation, etc., and a URL or DOI. Use the style used in the IEEE Transactions on Image Processing. Where possible, cite sources in high-quality peer-reviewed journals (e.g., IEEE Trans. Im. Proc.) or highly selective conferences (like CVPR) rather than random preprints and tech reports on the web. When you access a paper on arXiv, check first for an officially published journal version and cite that instead of the arXiv version when available.
• Use Overleaf for LaTeX-style collaborative editing with a convenient interface. We will provide Overleaf templates for each project component (proposal, progress report, final report). For future reference, any IEEE member can use the pro Overleaf tools by linking with your IEEE Collaboratec account. See the Overleaf quick guide for the settings.
• For preparing the presentation collaboratively, you could use LaTeX with its beamer class, or google slides (you can even enter symbols with LaTeX abbreviations) or PowerPoint (its slide-by-slide recording feature is pretty useful).

• Equations should be part of full sentences and should be punctuated. Do not start a sentence with an equation. (Look in IEEE T-IP for examples.) Here is an example:
  
  Notice how commas are used after the list 1,...,L , in the text, and after the first displayed equation. Then the final equation has a period after it because it ends the sentence.
• Replace "in order to" with just "to". (I see this a lot.)
• Punctuate Latin abbreviations properly, e.g., as I did in this sentence. Notice the two commas, i.e., one before and one after as in this sentence too.
• The Latin "et al." is short for "et alia" so it must have a period.
• Do not use contractions in technical writing: use "do not" instead of "don't".
• Learn the proper use of "which" and "that" per the Chicago Manual of Style. Also see oxford dictionaries and these tips.
• Use commas properly.
• When reporting numerical values, think about how many digits are meaningful. Very rarely are all the digits in "12.3456" statistically significant or practically meaningful in signal processing problems.
• Avoid the ambiguous word "performance" when describing methods and results. Use more specific terms like resolution, noise, accuracy, error, speed, etc.
• Use active voice so that it is clear that you did the work.
  • No: A simulation was performed to evaluate the error. (This is ambiguous passive voice.)
  • Yes: We performed a simulation to evaluate the error. (This is clear active voice.)
• Some people object to active voice because it can end up with a lot of "we" in the document which some people think of as "not scientific." I disagree: scientists are people and we should own the work we do.
  However, to reduce the use of "we" in your writing while still using active voice you can have document components be the actor.
  • No: It is shown in Fig. 1 that the RMSE is large. (Wordy passive voice.)
  • No: We show in Fig. 1 that the RMSE is large. (Active voice, but still unnecessarily wordy.)
  • Yes: Fig. 1 shows that the RMSE is large. (Most concise, and also active voice!)
  
  
  • No: The new algorithm is introduced in Section 2.
  • Yes: Section 2 introduces the new algorithm.
• Use IEEE style for equation references, i.e., "See (3)" instead of writing "See equation 3"

• It is easier to edit and track changes
  (e.g., on Overleaf)
  if you break the text into shorter phrases,
  like I did here.
  (Use more line breaks, like poetry!)
  Donald Knuth, the inventor of TeX,
  recommends this approach in his book.
• Use equation labels that describe the content like eq:model rather than using a number like eq:1, because when we reorganize the document the numbers will change but the content remains the same.
• Use \eqref{eq:model} instead of (\ref{eq:model}) because different journals have different equation formatting rules, like (1) or [1] or "eqn. 1", and this macro can adjust based on the style.
• Acronyms like "SNR" should not be in italics. (It is not a product of three variables S N R. Use \text{SNR} or \mathrm{SNR} in equations. Better yet, make a macro \SNR like this:
  \newcommand{\SNR}{\mathrm{SNR}\xspace}