Download (PDF, 9MB, Not barrier-free file.) - Nestor

submitted must be free from DRM or other security or protection measures (e.g. passwordprotection, etc.). Thus, for example, the record accessible athttp://hdl.handle.net/2128/3140 (25.10.2009) links to an encrypted campus presspublication protected against changes by password. As already explained above, this ishighly problematic if the document is to be preserved in a long term archive. Thus, itseems advisable not only to modify the policy accordingly, but also to take measures toascertain that the repository indeed has complete control over these digital objects asotherwise long-term preservation might not be possible or considerably more difficult torealized for these.Ingest: Quality AssuranceIntegrity: DSpace versions up to 1.5 carry out a format identification which, however,is merely based on the filename extension which is checked against a (locally installed)format registry. 105 Thus, as Larry Stone, commenting on problems with the current method,explains in a paper delivered at the Open Repositories Conference 2008,[t]he data model for technical metadata (the BitstreamFormat object) has been essentiallyunchanged from the inception of DSpace through release 1.5. Although its design anticipatedthe use of external format registries, it was never completed. There are some other problemswith the current data model and implementation:1. Formats are identified by arbitrarily-assigned descriptive names such as “Adobe PDF”which have no meaning outside of DSpace. This impedes any attempt to share formatdescriptions with any other application.2. There is no provision for collecting more extensive format technical metadata, such asstandards documents, that would help future preservationists interpret obsolete formats.3. A Bitstream's format is only identified by comparing its filename extension to entries in theformat registry. This method is prone to errors, ambiguous results, and outright failures [...].(2008, 2)Thus, currently neither a thorough format identification nor a format validation, whichwould have to be performed by an external tool such as JHOVE, are carried out forsubmissions to JUWEL. In consequence, the question of whether a file is functional or notis at the moment merely answered by means of opening it, which clearly is no sufficientprocedure if digital objects are to be preserved for the long term.Additionally, the DSpace submission workflow gives authors the opportunity todownload their publications immediately after upload in order to verify that the MD5checksum generated during the ingest process is the same as the one of the originalgemäß Absatz 1 zu verwerten” (Juwel Policy).105 Plans exist to implement an automatic format identification service into DSpace making use of externalformat registries as well as a new data model for technical metadata in a future DSpace release. Theseefforts are explicitly related to considerations concerning digital preservation. As Stone explains, “[t]oaddress digital preservation problems and take advantage of tools currently being developed, DSpaceneeds more fine-grained format classification, as well as globally-recognized identifiers for formats […].Both of these are provided by a preservation-minded external format registry such as PRONOM or GDFR”(Stone 2008, 5.2). In addition, the importance of format validation has been recognized, and according toStone the new data model “is ready to record validation; successfully validated Bitstreams are marked witha Confidence value of 'VALIDATED'” (Stone 2008, 5.2).As outlined in the DSpace wiki, future projects will address the following preservation applications:“Detecting and notifying administrators of obsolete formats in the archive,” “Format migration andnormalization (migration on ingest),” and “Data format validation (some of which is already implemented inpending JHOVE integration work” (DSpace Wiki 2008a).39