ausblenden:
Schlagwörter:
-
Zusammenfassung:
With the foundation of DataCite in 2009 and the technical infrastructure
installed in the last six years it has become very easy to create
citable dataset DOIs. Nowadays, dataset DOIs are increasingly accepted
and required by journals in reference lists of manuscripts. In addition,
DataCite provides usage statistics [1] of assigned DOIs and offers a
public search API to make research data count. By linking related
information to the data, they become more useful for future generations
of scientists. For this purpose, several identifier systems, as ISBN for
books, ISSN for journals, DOI for articles or related data, Orcid for
authors, and IGSN for physical samples can be attached to DOIs using the
DataCite metadata schema [2]. While these are good preconditions to
publish data, free and open solutions that help with the curation of
data, the publication of research data, and the assignment of DOIs in
one software seem to be rare. At GFZ Potsdam we built a modular software
stack that is made of several free and open software solutions and we
established 'GFZ Data Services'. 'GFZ Data Services' provides storage, a
metadata editor for publication and a facility to moderate minted DOIs.
All software solutions are connected through web APIs, which makes it
possible to reuse and integrate established software. Core component of
'GFZ Data Services' is an eSciDoc [3] middleware that is used as central
storage, and has been designed along the OAIS reference model for
digital preservation. Thus, data are stored in self-contained packages
that are made of binary file-based data and XML-based metadata. The
eSciDoc infrastructure provides access control to data and it is able to
handle half-open datasets, which is useful in embargo situations when a
subset of the research data are released after an adequate period. The
data exchange platform panMetaDocs [4] makes use of eSciDoc's REST API
to upload file-based data into eSciDoc and uses a metadata editor [5] to
annotate the files with metadata. The metadata editor has a
user-friendly interface with nominal lists, extensive explanations, and
an interactive mapping tool to provide assistance to scientists
describing the data. It is possible to deposit metadata templates to
fill certain fields with default values. The metadata editor generates
metadata in the schemas ISO19139, NASA GCMD DIF, and DataCite and could
be extended for other schemas. panMetaDocs is able to mint dataset DOIs
through DOIDB, which is our component to moderate dataset DOIs issued
through 'GFZ Data Services'. DOIDB accepts metadata in the schemas
ISO19139, DIF, and DataCite. In addition, DOIDB provides an OAI-PMH
interface to disseminate all deposited metadata to data portals. The
presentation of datasets on DOI landing pages is done though XSLT
stylesheet transformation of the XML-based metadata. The landing pages
have been designed to meet needs of scientists. We are able to render
the metadata to different layouts. Furthermore, additional information
about datasets and publications is assembled into the webpage by
querying public databases on the internet. The work presented here will
focus on technical details of the software stack. [1]
http://stats.datacite.org [2]
http://www.dlib.org/dlib/january11/starr/01starr.html [3]
http://www.escidoc.org [4] http://panmetadocs.sf.net [5]
http://github.com/ulbricht
