Reference

Prior to a data import, the reference related to the data has to be defined in the reference table. If a paper has a persistent identifier (DOI), it is entered in the URI-field; if no identifier exists, a link to the paper in pdf-format might be given instead (see the pdf-article on how to prepare and link pdf-files for Pangaea).

To add a new reference in 4D, click on Reference, click new. As a first step (if known), you should enter:


 * URI: should be a DOI, urn:nbn, hdl etc. or any other persistent identifier. NO links to web pages - use persistent identifier when ever available! (add the 'protocol' to the URI with ':', e.g.
 * pdf-files can be stored in the repository ePIC which provides a handle for persistent linking or any other repository with persistent identification.
 * pdf-files can be stored in the repository ePIC which provides a handle for persistent linking or any other repository with persistent identification.
 * pdf-files can be stored in the repository ePIC which provides a handle for persistent linking or any other repository with persistent identification.

You can now click on "Get Metadata" which will try to extract all of the following information from the given URI. This works with DOI Names (provided by Crossref or DataCite), some Handle types (only if provided by ePIC), but also with plain URLs to PDF/HTML/... files. For DOIs the metadata is requested from the publisher through web services. Otherwise it is extracted from the ePIC website for handles pointing to ePIC or, as a last chance, extracted directly from the PDF or HTML page (title, sometimes also authors and year) using Apache TIKA (see the webpage for all supported file formats).

It may happen that some of the information below needs to be completed, e.g., if the authors do not yet have entries in the staff table:


 * Author(s): Names must already be defined in the staff table and will be added by multiple choice selection
 * Title: full title; no abbreviations; no full stop at end
 * Source: this field should be used only for any non-standard unique source information, e.g.
 * Books - In: Bleil, U & Thiede, J (eds.), Geological History of the Polar Oceans, NATO ASI Series, Kluwer Academic Publishers, London
 * grey literature like special reports etc.
 * Journal: relational field to a table with 16500 most common journals
 * Volume: 53(2); set to 'submitted' or 'in press' if not yet published
 * Pages: 34-67 or 123 pp for complete books, thesis, reports etc.
 * Year: 1998 (for papers 'submitted' or 'in press' leave empty)
 * Serial No (optional): proprietary publication number, e.g. AWI publ No 1234 or PKDB56789
 * Type: choose as appropriate from the choice list
 * Keywords: select from thesaurus or define new as required

If using copy/paste to enter new references, avoid to add carriage returns or redundant spaces.

''In the future there will also be a service to copy/paste a full formatted citation (with or without DOI) into the reference window and look up the DOI automatically. Until this is implemented, use the simple text query interface provided by CrossRef as a workaround.''

Attention is required when a reference is deleted! Even if relations to data sets exist, the reference can be deleted. While pressing  the curator is prompted with a requester: x related records in 'DATA SET' - do you want to delete relations?  will delete reference and relations.

A list of references can be imported as tab-delimited ASCII-file. A format and header template is provided in the 'IMPORT FORM FOR REFERENCES' by using Menue File/Import, choose References. Authors are also recognized by staff IDs. Author in the import list will be recognized if:
 * name and first name are identical, notification if several similar names
 * name and initials are identical, notification if several similar
 * name identical and initial identical with initial of first name (first letters of several first names are added for comparison)
 * name identical and 1. initial identical with initial of 1. first name (only if no other similar names are archived, e.g. Shackelton, N is recognized because Shackelton, Nicholas J is the only record with name Shackelton in the staff list.

For the reference management software EndNote X an export style is provided:.

During data import, the relation between a data set and its reference is set. Several references can be related to a data set but in principle, real primary data should be supplementary to one reference only.

Supplementary data should be submitted and imported before a paper is in press. After import and proofread of the data, the author is provided with a DOI for data citation/identification to be used in the publication. The paper should not just refer to the Pangaea domain, but give the link precisely pointing to the relevant data.

Examples:
 * as a link to one specific data set or
 * as a link to a number of data sets (childs), grouped in a parent set or
 * http://www.pangaea.de/search?q=@Ref896 (896 is the PANGAEA internal reference ID) to show many data sets related to a publication without a parent definition. This link may be used if many data sets (some thousand) are related to one reference.

For refering in a publication to its primary data, the author should refer to the data by using a note like: Supplementary data are available at 

Links

 * Connotea Free online reference management (provided by Nature)
 * refbase Web Reference Database
 * example: Literature Database SK Ozeanographie
 * ISO standard for bibliographic referencing