Paper with lots of Greek letters

Question

I developed a system, and wrote a paper about it. It includes formulas and pseudo-code. During the course of development I used Roman letters in the formulas. I eventually ran out of meaningful letters (letters that I could associate to some extent with what they represented), and started using Greek letters. The development went on until somehow I ended up having about 8 Greek letters and 2 Roman letters used in the final formulas and pseudo-code.

Now that the paper is written, I wonder if having all of these Greek letters is unusual. Would it be better to replace them with Roman letters? Would it be better to use Roman letters with subscripts, such as t for time, and t_scaled for scaled time, instead of using Greek letters for that?

EDIT: To clear things up a little, my general question is: is it preferable to use Roman letters or Greek letters in a technical (engineering) publication, and is it acceptable or even preferable to use letters with word subscripts instead of entirely new symbols to label related variables or constants?

Answer 1

This depends entirely on your field, and specifically, what the norm is in your field. In mathematical fields like statistics, economics, physics, and math itself, and to a slightly lesser extent in chemistry and computer science, using Greek letters as variables is exceedingly common - in fact, it's unusual to see a paper that doesn't use a few, and there are many quantities for which the well-established standard notation involves a Greek letter. In other fields, perhaps that's not the case.

If you're still not sure after examining the norms in your field, I'd just go ahead and do it. Given that mathematicians use Greek letters all the time, it stands to reason that when you are doing math in another field, you should be able to do so as well.

However, I would caution you about one thing: try to make your variable names descriptive. That is, it should be easy for the reader to make the association between the variable name and what it represents. So when there is a standard notation for some quantity, use that; when there isn't, subscripted symbols like "t_scaled" are one good way to do it. They do take up more space on the page, though, and can make formulas look unwieldy, so you have to strike a balance between aesthetics and clarity of meaning.

Answer 2

it preferable to use Roman letters of Greek letters in a technical (engineering) publication, and is it acceptable or even preferable to use letters with word subscripts instead of entirely new symbols to label related variable or constants?

In addition to the excellent answers that have already been given, I would like to stress the fact that your original question cannot be answered in the generality that you are apparently looking for. By and large, Greek letters are OK, as are Roman letters, as are either of those with sub- or superscripts. What is preferable depends entirely on context, the established conventions and notation in your field, and personal preferences.

When it comes to notation, the 3 golden rules are:

1. Make it intuitive. Don't use epsilon to mean "a really large value". Don't use pi to refer to a constant value of 8.5. If you, for example, have instances and timestamps, don't use "t" to refer to the instances and "i" to refer to timestamps (real-life example I recently reviewed).
2. Make it consistent. Use the same notation in the entire paper. Use the same notation patterns throughout the paper (if you use capital Roman letters for a constant, use capital Roman letters for all constants, etc.).
3. Make it minimal. Never introduce notation that you are not actually using.

Answer 3

The writing style is really something very individual, and may vary a lot. There is nothing wrong with Greek letters per se, but it is slightly concerning that you use them because you ran out of meaningful Roman letters. That could indicate that you are using really a lot of notation in your paper, and from a point of reader it might be difficult to keep in mind what do these letters mean.

The use of semantics, i.e. notation like n_cars or t_scaled, may really help to reduce the number of "independent" symbols and improve readability. You may also wish to structure the use of variables, e.g. use Greeks as integers and Roman as everything else. It is very good idea to stick to some classical notation worked out in textbooks and major articles in your area of research. Ultimately, you may wish to present a table of all your letters with their meaning after the first section of your paper.

Answer 4

Greek letters are generally fine. In fact, some things are conventionally reffered to by greek letters e.g standard deviation is sigma.

xLetix's rules are excellent advice. I would just add to please put explainations of your variables close to where you use them. There are two things people do that annoy me in this respect.

• Putting a glossary at the beginning of the text and then don't explain there symbols. A glossary itself is fine but please put explainations in text as well as I don't want to have to flick back to the start to know what every symbol means.

• Similarly, people often put an explaination when they first use the symbol (often in the introduction) and then use it again several pages later with no explaination. This is fine if you have a few obvious symbols but when many with unclear names are used it is difficult to keep track of them. Just writing the name in words generally suffices.