When DSpace was initially built, the assumption was that "unpublished" documents (e.g. gray literature in general) should be auto-assigned a "dc.date.issued" if it was unspecified.
This logic is OK if your repository is primarily gray literature.
However, if your repository contains some documents which have been issued before, it's more accurate to either force the user to assign the "issued date" or leave it unassigned.
Google Scholar has reported (see
DS-1481) that DSpace's tendency to auto-assign this date has led to problems in Google Scholar's ability to index DSpace content. It also means that DSpace content can become harder to locate in Google Scholar when the issued date assigned is highly inaccurate.
I'm working on a Pull Request that does several things:
1) DSpace will no longer auto-assign a "dc.date.issued=[today]" when the field is missing (except where described in the next few cases)
2) However, when "Initial Questions" are enabled from the UI, DSpace will still assume "dc.date.issued=[today]" when a user says a work is "unpublished". This allows some form of backwards compatibility. Institutions who have a lot of gray literature in DSpace can enable "Initial Questions" for one or more Collections to have DSpace auto-assign issued dates again. (Please note that "Initial Questions" are disabled by default in 4.0, see
3) Adds a new option (for SWORD or Bulk Ingest tools, etc) to pass in the literal string "today" as the "dc.date.issued" to tell DSpace to set 'dc.date.issued=dc.date.accessioned". This means that you can now decide (in your import metadata) whether you want DSpace to auto-assign issued dates when performing a bulk import.