For many data classes (e.g. genes, labs, etc.), submitters should be able to specify the publications associated with that data item. This process must be as streamlined as possible, because users have very little patience with this.
A crucial issue is that publication data that are somehow malformed (e.g. the submitter didn’t type the date in the correct format) can essentially be lost in the database, because later queries will not be able to retrieve them. For this reason, all new PUBLICATION records are entered by the DB admin. Specifically, the way it works is as follows:
1) When asked to list publications, users provide the name of the first author and the first n words of the title for each publication they wish to list.
2) Upon submission, the system matches this list against the PUBLICATION database.
3) All publications it finds in the database are added to the current data record (e.g. an image file being submitted).
4) For any publications it can not find in the database, the user is prompted to send email to DB admin. asking that (a) publication record be created and (b) added to the current data record. For the present, nothing is attached to the data record --> its publication list remains without a temporary record. It is crucial for the DB Admin to process requests to enter new or missing publications promptly. Upon receiving the request, the DB Admin enters the new PUBLICATION record and adds it to the appropriate data record.
5) If there are any problems creating the new publication record, the DB admin corresponds by email with the submitter to iron them out.
Note: It is anticipated that these explicit requests for data records to be added will be quite infrequent. Hopefully, most relevant publications will be automatically added to the database as they are published.