I am writing a computer science paper in which I cite several lengthy sources (text books and standard specifications). Whenever I cite such a source, I add the page numbers of the subsection to which I refer, e.g.:

Bla bla is considered very important [52, pp. 210-214].

One of the reviewers of the paper remarks the following:

Do not use references with indication of page numbers (e.g., [52, pp. 210-214]), this is very uncommon.

Is this really true (in computer science)?

Am I wrong to include the page numbers in the reference?

The reference in question is a 400+ page book that broadly covers a whole subject area and I thought knowing which pages I actually refer to might help the reader. There are several other (and longer) sources, such as technical standards, which I cite in the same fashion.

  • 20
    I can't address how common citing page numbers in books is in computer science, but it certainly ought to be standard. I'd recommend ignoring the reviewer's comment. – Anonymous Mathematician Oct 23 '14 at 12:47
  • 15
    Yes, it's uncommon; no, you're not wrong to do it. – ff524 Oct 23 '14 at 12:59
  • 1
    It seems like a practical issue as well: Many CS textbooks that I've read were digital copies with no page numbers at all! – Brian S Oct 23 '14 at 16:18
  • 5
    This is uncommon for a reason. You should not reference by page number unless there is really no better way to reference the part you need (which means that the book is badly organized). The book might grow by a couple pages in a new edition and your page-number references are screwed. – darij grinberg Oct 23 '14 at 16:24
  • 2
    @darijgrinberg: Based on the answers, I'm not sure this question is particularly asking for page numbers as such, rather than for any kind of further specification where to find the particular cited statement in the referenced work. – O. R. Mapper Oct 23 '14 at 17:20

Many computer science venues are fairly picky about their citation styles and also use highly abbreviated citation styles that don't give any option for including additional information like a page number.

If page numbers are critical to your citation, then you have two options for how to include them while conforming to the required style:

  1. If you only use one narrow set of pages in the book, use an "in-book" style citation (@inbook in BibTeX) that includes the page numbers in the reference.

  2. Alternately, or if you refer to more than one set of pages, you can include the pages in the text rather than the citation:

Bla bla is considered very important, as described on page 210--214 of [52].

  • 2
    ACM, for example, does not include page numbers in the in-text reference style. – Bill Barth Oct 23 '14 at 13:10
  • 1
    Thanks @jakebeal and @bill-barth! This seems to be a good solution for my problem. Interestingly, the IEEE style guide uses the "in-book" style, but also [52, pp. 210-214] (see page 7). – joergl Oct 23 '14 at 13:23
  • 2
    @JukkaSuomela I've happily used inbook for a long time. Perhaps you can provide a reason that it should not be used, beyond the ad populum fallacy in the linked post? – jakebeal Oct 24 '14 at 1:32
  • 1
    "The best way to conform to computer science citation styles is to use the LaTeX packages" - when relying on that rule, contrary to what @BillBarth suggested, all three LaTeX packages from the linked ACM site happily display page numbers in the in-text reference style, when citing in the \cite[page]{source} style. – O. R. Mapper Oct 25 '14 at 22:50
  • 1
    @O.R.Mapper Awesome. Then I would totally trust ACM's machine-interpretable de facto standard over their human-readable and ambiguous documentation. – jakebeal Oct 25 '14 at 22:52

Specifying which part of a long book or article you are referring to in your citation is a sound practice, but in general section, theorem or equation numbers are more common than page numbers for this task: [52, Section 3.5] or [52, Equation 3.23] rather than [52, Page 132-134].

Personally I use page numbers only as a last resource where nothing else is available: that is, when the section or equation I want to refer to does not have a number of its own.

I do not have a compelling reason for this, if not that section numbers are more robust to minor pagination changes, but this is the behaviour that I have seen more commonly.

  • Thanks, I aggree that section numbers or equations are more stable than page numbers. I think I will combine your solution with jakebeal's, e.g. "...as described in Sect. 3.1 of [52]." – joergl Oct 23 '14 at 13:26
  • 4
    Why are section numbers more stable than page numbers? For a formally published work (i.e., not on ArXiv), neither should ever change, except by the publication of a new edition of a book. But, in new editions, it's very common for section and theorem numbers to change, too. – David Richerby Oct 23 '14 at 16:40
  • 2
    @DavidRicherby: in a preprint, the pages are numbered 1, 2, 3, ... but in the published paper they are numbered in sequence with all the other papers, so the pages might be 140, 141, 142, ... The section numbers will be the same, however. So in fields where preprints are common, citing sections or theorem numbers has practical advantages. – Oswald Veblen Oct 24 '14 at 1:03
  • 1
    @DavidRicherby: They could be in principle, but the page numbers are essentially guaranteed to be different, while the section numbers are relatively likely to be the same, in my experience. Things would be different in fields where they don't release preprints. – Oswald Veblen Oct 24 '14 at 13:37
  • 2
    Just to be clear, I am not saying "you should cite section numbers from an arxiv version without checking that the ones on the published version are the same"; I am saying "you should cite section numbers from the published version, not page numbers, so that in 99% of the cases a reader without access to the published version can use the arxiv one". – Federico Poloni Oct 24 '14 at 13:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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