How to address dependencies when publishing a chain of papers

Question

In computer systems, it's relatively common to have a large problem which is broken up into multiple sub-problems, each of which is substantial in its own right and is in the realm of different subsets of students. Suppose there's two parts, A and B. Suppose that properly presenting B requires explanations and background which A establishes (concepts, principles, etc.), and preferably would reproduce some figures for clarity. However, A has been written but not yet accepted/published.

How should the authors of B proceed? Ideally the figures would be reproduced from the drafts and clearly cited as copies, but it seems weird as the original is yet unpublished. This also creates the worry that reviewers for a later submission of A would complain that these contributions are not novel as B's earlier publication have already presented them.

Answer 1

Dependencies like these are a great place to use preprints.

The arXiv is usually a good preprint server for computer science research (other fields use others, like bioRxiv, ChemRxiv, etc). Your preprint can go up immediately, and then it's publicly available with a DOI for citation ever after.

This both clearly establishes ordering and allows paper B to cite paper A in its preprint form. Even if paper B makes it all the way to publication before paper A, the preprint can be linked to the final form after paper A is published, which indexing and citation services use to consolidate the versions together.

Finally, most journals allow preprints these days, and some even force you to use them, so in most circumstances there is no downside to posting a preprint.

Answer 2

Preprints, as pointed out by @jakebeal, are a very good way to deal with your problem.

However, you might want to avoid making A publicly available for some reason before submitting B for publication. In this case, it is often possible to submit additional material (A in your case) that is necessary for review but not meant for publication with B. In B, you can then include temporary citations pointing to A that you have to replace before publication of B. You can additionally explain your situation to the editor in the cover letter. Then the reviewers have all the information they need to assess B.

There is one obvious drawback of this approach: You might be unlucky that B is accepted for publication before A, and you have no proper way to cite A. I have seen quite a few times, especially in older papers, that another paper of the same authors is cited as "in preparation" or "submitted for publication". I cannot say that I would recommend this (it is quite frustrating if that very paper never appeared), and fortunately it seems to be uncommon nowadays.

Another strategy is obvious: Wait with submitting B until A is published.