How to reference original paper in expanded journal version

Question

Many journals allow to submit an extended version of a conference paper as, for example, Springer:

Expanded versions of papers which have previously appeared in conference proceedings will be considered and this request must be made at the time of submission; the authors must clearly state the conference publication and provide a statement of revisions/updates that have been made to the journal paper.

When writing such an extended version how do you reference the conference paper? I can imaging several possibilities:

1. Treat it like every other paper in related work
2. Ignore the conference paper completely and present all results as if for the first time.
3. Clearly indicate that you wrote the conference paper as "... in [1] we showed that and shortly review the approach ..."

What is correct way to write an extended journal (in CS)?

Answer 1

You write the journal version so that it is entirely self-contained, as if all results were new (regardless of whether they were already present in the conference version).

Then, in appropriate places, you refer to the conference version and explain how the current version is different. For example, you could have a footnote on the first page or an acknowledgement in the end of the paper, along these lines:

This is an extended and revised version of a preliminary conference report that was presented in FOO 2015 [1].

Here [1] is a reference to the conference version, and you include it in the list of references as usual. There is no need to cite it elsewhere.

If it makes sense, you can also discuss the difference between this version and the conference version in more detail e.g. in the "Related work" section. This might be a good idea especially if the journal version is not just an extension but some results are fundamentally different. However, in straightforward cases this is not usually necessary.

---

(Please note that this answer is specific to computer science, which has its own peculiar approach to conference publishing. All other fields are different.)

Answer 2

I cannot say about the CS field specifically, but, in general, I would just provide a note (either at the beginning, or at the end of the paper), saying something along these lines:

"The present paper is an extended version of our paper, presented at ... conference [1]. This paper significantly expands ... and ...".

Answer 3

In CS particularly, there is usually a step during the paper submission process in which you are required to disclose whether a part of your paper has been published elsewhere (or some similar phrasing). There you list the paper(s) your current paper expands.

Further, you state in the cover letter the circumstances in which your new paper was created (e.g. "in the conference proceedings, we show that [...], this approach is improved in this paper by [...], yielding [...] results. for better comparison, a new experimental setup was used, so the results of the conference paper are rerun in the new environment, which affected the conclusions [...]" - you get the idea). In other words, you state in a few paragraphs (keep it short though, 1-2 paragraphs should be enough) what's new in you paper compared with the previously published one and how is that significant.

As for your paper, you use your conference paper as any other reference (including the related work review), but a clear emphasis on the relation between the papers should be made (e.g. in the related work section). The style depends, you can be personal or not, it doesn't really matter. I stick to the same style I use for other references, i.e. I don't explicitly state that the cited paper is written by me.