In a recent question on how to deal with submitting several papers on similar subjects at the same time, one of the (good) practices suggested by the answers is make sure that they cite each other, and that each result appears as original only in one of them.
One may want to put informative references from one to the other; for example:
- paper A "An application of this result is in paper B";
- paper B: "We apply this result from paper A".
From the point of view of providing all the relevant context to the reader, this might be the most informative thing to do. However, the other use of citations is for computing metrics, and in this respect this practice could be seen as a dishonest way to improve one's citation count, since you get two (self-)citations instead of the one you'd normally get if the two paper were not written simultaneously.
Hence, the question in the title: Is this practice recommendable and/or acceptable? Or should I decide an implicit temporal order for my simultaneous papers, and only cite older ones from newer ones?