Imagine you are a peer reviewer presented with a paper that states an algorithm was implemented in python to test an analytical process on large datasets. The details of how the algorithmic process are detailed extensively via mathematical denotation and formulas, but no actual pseudocode or programming examples are shown. Would you consider the paper to be incomplete, or would the mathematical examples suffice (provided they are robust enough and don't withhold any information whatsoever)?

  • 4
    I expect it depends on the requirements of the journal that the paper is submitted to.
    – astronat
    Dec 28 '20 at 9:17
  • 1
    Entirely depends on the target venue (journal or conference) and the preferences and expectations of their reviewers. In a programming-focused venue, you might not get the reviewer feedback "rejected due to lack of pseudocode", but reviewers might struggle to understand your algorithm and therefore give less favorable reviews. Dec 28 '20 at 9:37
  • You could add in a footnote the URL of some implementation e.g. on github.com Dec 28 '20 at 9:59

In theoretical computer science (I cannot speak for other fields) it is very common not to give pseudocode, but only a high-level description of the algorithm. Your paper should always explain your ideas as clearly as possible, and as an author it is your job to select the most suitable presentation. This does not necessarily need to be pseudocode. Painstakingly presenting every little implementation detail may not add much to the paper and instead distract the reader. If you think foregoing pseudocode results in a better paper, by all means do so (unless the journal you are submitting to specifically requires it).

If your paper presents an implementation in Python, I would expect the source code to be publicly available somewhere and referenced in the paper.

Not the answer you're looking for? Browse other questions tagged or ask your own question.