Should papers try to be self-contained?

Question

I gave my supervisor a paper draft today. They told me that I need to cut out the 5 pages of appendices that give all the equations (along with references of their original sources) needed for reproducing the results found in the main text.

My supervisor's argument was that papers should not be self-contained but rather that papers should direct readers to the different sources that came up with these equations. My supervisor went on to say that others might think I am trying to steal their future citations by attempting to have people cite my paper for a formula that someone else came up with (even though I provide attribution for the formula).

In my particular case, I am submitting to a journal with no page limits, albeit there are page charges that my supervisor will pay for.

Obviously, I am going to do what my advisor says for this paper. However, is this the general advice that I should carry with me throughout my career?

Answer 1

There is no hard and fast rule on whether to include calculations in papers that have appeared in other papers. It is a judgment call on the part of the author. For my papers, I have sometimes included these equations (either in the main body of the paper or in the appendix) sometimes I have omitted them and directed the reader to the bibliography, and sometimes I have just provided a brief sketch of the argument.

The reason to include them is because sometimes they are convenient for the reader. The main reason to exclude them are because they detract from the focus of the paper. As an extreme example, if your important new ideas in a paper take up 3-4 pages, it is strange to include 50 pages of previously published calculations, say, even if it is in an Appendix. It makes it harder for me as a reader to understand exactly what is the main point of your paper, and I might get the impression that the authors are adding fluff to make their contribution seem more important than it is. You want to keep your novel, important ideas at the forefront of your paper.

And don't discount the page length: even if they are no explicit page limits in a journal, you generally want to convey your idea in as few pages as possible. First of all, is the possibility that potential readers might be discouraged from reading your paper if it seems too long. Secondly, journals do hold longer articles to higher standards. When I was submitting to the top ranked journal in my field (which had no explicit page limits) we were told by a senior professor that our idea was worth publishing there if we could keep it under 30 pages, and probably not worth publishing there if it hit 40. In other words, there was a sense where our idea was not worth wasting 40 pages worth of someone's time, although it was worth wasting 30 pages worth of someone's time.

In summary, papers don't necessarily have to be self-contained, and how much to include previous work is a judgment call on the part of the author. I will add that as a student, it is probably best to defer to the advisor when making these judgment calls, as the advisor will be more familiar with the conventions and culture of your research area.

Answer 2

I think it's simple: if the tools or techniques that you need are easily accessible to the reader with clear references etc., then I think the best you can do for the journal and readers of your work is to give the intuition of that work and a clear reference. Nothing more is needed and indeed sometimes full details of such techniques might produce opposite than intended effects.

Keep your paper simple and to the point and avoid "fillers".

Be also careful that many journals with no page limits actually do have such limits: Siam journal on Discrete Maths is one on my mind in my field at least which, although it has no explicit page limit, it silently imposes higher standards to longer papers.

Answer 3

Papers must be self-contained to the point where, given relevant background knowledge and access to the same equipment/tools, a secondary party is capable of attempting to reproduce and/or verify the results of a respective paper. This is not about achieving brevity, but about ensuring clarity and comprehensibility. Detailed methodologies, data sets, and analysis techniques used in the study should be provided. The aim is to encourage transparency and facilitate scientific discourse by making the paper as accessible and understandable as possible, rather than merely limiting its length.

Answer 4

I almost always make them self-contained. Technically there were just 3 exceptions during my whole career where I needed to "borrow" some big chunk of an established theory well-written in some easily accessible textbook that would be hard to compress into a short sequence of relevant lemmas needed for getting the result I need.

In my opinion, the purpose of writing a mathematical paper is acquainting the reader with either new result or a new (at least, for him/her) idea (the latter IMHO is more valuable, but much more rare as well) in the most efficient and the least painstaking way. Everything that goes towards this purpose should be welcome and everything that goes against it should be frowned upon.

Creating a long chain (if it were just a chain, usually it is a branching tree growing exponentially in size when you try to dig deeper and deeper!) of references, IMHO, impedes the process of reading and makes the verification of the proof extremely complicated. You are no longer saying: "Here is the result and here is how to obtain it in the shortest way from the basic principles", thus putting it in the context of not the irrelevant temporal history of who did what and when or what is the latest fleeting fad of the times, but in that of what exactly was really needed to be developed before your approach/technique/observation became possible as it is. Instead you downgrade your exposition to the conditional statement that "if 5 references I give, 25 references contained in those 5, 125 references contained in those 25, etc., are all correct, then what I'm writing follows" and create the (most often false) impression that to understand your work entirely, one needs to read several thick textbooks, several hard to get research papers often written in not the clearest possible way and proving only something similar to what you really need in a completely different notation and with a completely different emphasis, so you still need to make the final adjustment yourself. A 30-40 page argument written in a clear and coherent style becomes a monster encompassing thousands of pages of patchwork of loosely hanging fragments pulling in various directions and making the reader totally confused about what the main thing in the whole business is (note that the main thing in the proof is not necessarily the last step you added to it, even if the latter is truly novel).

That is about references the theorems. For just the equations, you can do it if all equations you need are contained in one textbook that is both well-written and easy to access and derives them sufficiently early in the exposition (so you don't need to read 300 pages of the linear text or do the painstaking backward search trying to figure out what on those 300 pages is relevant to what you are after to arrive at the result). If your advisor can come up with such a textbook, cite it. If he cannot, I would recommend denying his request for the appendix cut and keep the things as they are.

Now about My supervisor went on to say that others might think I am trying to steal their future citations by attempting to have people cite my paper for a formula that someone else came up with. Are you really "trying to steal"? If you aren't, don't bother too much about what other people might think (or even do think) about you. I'll tell you one unflattering truth: most of them do not think about you at all and are completely unaware of your existence and totally indifferent to it. When they are reading your work, they are interested in what is written there, not in you personally. Their goal is to understand your writing, not to judge your moral character based on it.

As to In my particular case, I am submitting to a journal with no page limits, albeit there are page charges that my supervisor will pay for., I believe that it explains the whole thing in a crystal clear way: your supervisor wants to spare a few bucks and, since people are usually reluctant to say such things plainly, invents a whole bunch of bogus "ethical" and other explanations of why the cut should be made. I have no idea about what moved you to the submission to a predatory journal in the first place (every journal that charges authors for their own work is predatory in my classification) while there are plenty of decent ones around (that is on the topic of "advice that I should carry with me throughout my career"), but once you've made such choice, you created an obvious conflict of interest, which you are trying to deal with now. In general remember that you don't need to publish anything beyond the initial minimal amount of few papers to get a decent position and then occasional couple of papers for promotion and such. You know how to prove the results you proved. It is the outside world that doesn't, and if it tries to make you pay money for sharing the proofs with it and going through the long painstaking process of creating the clear exposition in the form of "publication fees", "conference registration fees" and other modern nonsense, you are always free to send it to the Hell and keep the results to yourself, just sharing them with a few friends and colleagues in the form of private communication.

Just my 2 cents, as usual.