I would argue that datasets uploaded to a GBG are publications.
Researchers in the taxa concerned may not only wish to cite such publications but additionally the locating of such publicly available data may not only benefit their research but also raise the profile of the GBG (be it Gateway, Atlas or GBIF).
Accordingly it should be made possible for our uploads to be referenced as a publication.
I've attempted to treat a Gateway upload as a publication by adding it to ResearchGate. It keeps asking me to do things which aren't easy to comply with, for example it asks me "does the journal you published in support self-archiving?" and then requesting that I upload the full text of the publication. All well and good until you try that from the General tab of a dataset page (which seems to summarise the dataset sufficiently adequately for the purposes of a researcher) because there's no sensible way of obtaining a file from that page that fulfils their requirements to dump a file (Add full text).
Bridging this gap between GBG data and custom and practise in the world of research and published papers is urgently needed. I've come across many papers where the authors are clearly unaware of GBG data and think everything's to be found in museum collections - 2 centuries of naturalist's efforts ignored.
Give us DOIs and a means of turning our uploaded dataset metadata page into appropriately citable documents please.
{And a new section for the Atlas in this Forum)