Requirements for SIP structure

CSIPSTR are requirements according to the E-ARK specifications (E-ARK (C)SIP specification (v2.2.0)). Those marked NBSIPSTR are requirements set by the National Library of Norway. The NBSIPSTR requirements are mainly stricter versions of the CSIP requirements in the E-ARK specifications.

Any information package submitted to the National Library of Norway is required to fulfill all CSIPSTR requirements as well as all NBSIPSTR requirements.

IDDescriptionRequirementCardinality
CSIPSTR1Any Information Package MUST be included within a single physical root folder (known as the “Information Package root folder”). For packages contained in an archive format (e.g. ZIP, TAR), the archive MUST unpack to a single root folder, see CSIPSTR3 og NBSIPSTR3.MUST1..1
NBSIPSTR1The Information Package MUST describe exactly one intellectual entity/resource. For more information see Intellectual scope of SIPs.MUST
NBSIPSTR2The name of the Information Package’s root folder MUST be the same as the @OBJID attribute specified in the METS.xml file located in the root folder. The name MUST only contain permitted characters, which are: ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz0123456789-_ . The name of the root folder SHOULD be a unique ID (e.g., a URN).

This requirement is a stricter version of CSIPSTR2.
MUST
NBSIPSTR3The Information Package MAY be contained in an archive/compressed form for transfer. If the Information Package is transferred in such a format, only TAR or ZIP formats MUST be used. If the Information Package is transferred in multiple parts, each part MUST NOT exceed 5 GB. See also CSIPSTR1.

This requirement is a stricter version of CSIPSTR3.
MAY
NBSIPSTR4root_folder/METS.xml

The Information Package root folder MUST include a file named METS.xml. This file MUST contain metadata that identifies the package, provide a high-level package description, and describe its structure, including pointers to constituent representations. METS.xml MUST meet the requirements specified in the METS.xml requirements.

This requirement is a stricter version of CSIPSTR4.
MUST1..1
NBSIPSTR5root_folder/metadata

The Information Package root folder MUST have a single folder named metadata. Metadata describing the entire information package belongs in this folder.

This requirement is a stricter version of CSIPSTR5.
MUST1..1
NBSIPSTR6root_folder/metadata/preservation

If preservation metadata is available, it MUST be placed in a subfolder called preservation. The E-ARK specifications recommend formatting preservation metadata using PREMIS (see E-ARK CSIP | Common Specification for Information Packages and CS PREMIS).
Preservation metadata refers to events and the processing history of the digital resource. This includes documentation of events such as file migrations to new storage environments, file format conversions, and the validation or verification of files.

This requirement is a stricter version of CSIPSTR6.
See also the requirements for using METS.xml for metadata in CSIP31-57.
MUST0..1
NBSIPSTR7root_folder/metadata/descriptive

The metadata folder at the root level MUST contain exactly one subfolder named descriptive. Descriptive metadata shall only be placed in this subfolder, and not within the individual representations. Descriptive metadata refers to information that describes the intellectual content of a digital object, such as an identifier, title, creator, and date of publication.

This requirement is a stricter version of CSIPSTR7.
See also NBSIP8-10 in METS.xml (SIP).
MUST1..1
NBSIPSTR8The descriptive metadata files in the information package MUST be formatted as plain text (UTF-8) (e.g., XML, JSON, TXT) and SHOULD follow a well-known metadata standard (e.g., DublinCore, MODS, MARC, EBUCORE, EAD, etc.)MUST
NBSIPSTR9root_folder/metadata/descriptive/

The subfolder descriptive MUST contain one or more files with descriptive metadata. For example, exports of catalog records, information from archival/management systems, and/or other forms of descriptions of the resource. This MAY be in the form of proprietary data from a given system, but SHOULD also be mapped to a well-known standard in accordance with the requirement in NBSIPSTR8.
MUST1..n
CSIPSTR8root_folder/metadata/other

If any other metadata are available, they MAY be included in separate sub-folders, for example an additional folder named other.
MAY0..n
NBSIPSTR10root_folder/representations

The Information Package MUST contain exactly one folder named representations. See Representation types for more information.

This requirement is a stricter version of CSIPSTR9.
MUST1..1
NBSIPSTR11root_folder/representations/primary_YYYYMMDD

The representations folder MUST contain exactly one representation named primary_YYYYMMDD (where primary is followed by the date the folder was created). The contents of this folder constitute the digital object that the package is intended to describe. This is the primary representation in the package (see Representation types for more information).

This requirement is a stricter version of CSIPSTR10.
MUST1..1
NBSIPSTR12root_folder/representations/

If needed, the representations folder MAY contain multiple representations. Naming of representations should consist of a meaningful name and the date the folder was created (meaningfulname_YYYYMMDD) (see examples under the pages for media-specific SIP requirements/profiles). Creating multiple representations is only relevant if it is considered necessary to preserve both an original primary representation and something derived from it. This may include an access copy, a processed, normalized, format-migrated, or repaired version of the primary representation.

The purpose of the derivative, how it was created, and its relationship to the primary representation SHOULD be documented in PREMIS events (CS PREMIS ). We discourage preserving derivatives that can easily and automatically be reproduced from the primary representation. See the principle of sustainability in digital preservation for more information.

This requirement is a stricter version of CSIPSTR10.
If the package contains multiple representations, they must follow the same structural format as primary_YYYYMMDD (NBSIPSTR13-17 and CSIPSTR13).
MAY0..n
NBSIPSTR13root_folder/representations/primary_YYYYMMDD/data/

Each representation folder MUST contain exactly one folder named data, which in turn MUST contain the data that constitutes the specific representation.

This requirement is a stricter version of CSIPSTR11.
MUST1..1
NBSIPSTR14root_folder/representations/primary_YYYYMMDD/METS.xml

Each representation folder MUST contain a metadata file named METS.xml, which includes information about the identity and structure of the representation and its components. METS.xml MUST comply with the requirements specified in the METS.xml requirements.

This requirement is a stricter version of CSIPSTR12.
MUST1..1
CSIPSTR13root_folder/representations/primary_YYYYMMDD/metadata

The representation folder SHOULD include a sub-folder named metadata, which MAY include metadata about the specific representation.
SHOULD0..1
NBSIPSTR15root_folder/representations/primary_YYYYMMDD/metdata/preservation

The subfolder named metadata within the representation folder MAY contain a subfolder named preservation. This is where preservation metadata specific to the representation should be placed. The E-ARK specifications recommend formatting preservation metadata using PREMIS (see E-ARK CSIP | Common Specification for Information Packages and CS PREMIS).
MAY0..1
NBSIPSTR16root_folder/representations/primary_YYYYMMDD/metdata/technical/

If technical metadata exists, it SHOULD be placed in the metadata subfolder, in a folder named technical. Technical metadata MUST be organized into subfolders named according to the type of technical information they contain. Technical metadata documents the technical characteristics of a digital file or resource, such as file format, resolution, compression type, dimensions, color depth, and software/hardware requirements for opening the file.

Examples of such files include metadata files produced by MediaInfo, ExifTool, JHOVE, etc. These can be placed in separate subfolders (e.g., technical/mediainfo, technical/exiftool, technical/jhove, etc.). (technical/mediainfo, technical/exiftool, technical/jhove, etc).

See also NBSIP20-27 in METS.xml (SIP).
SHOULD0..n
NBSIPSTR17root_folder/representations/primary_YYYYMMDD/metadata/source

If the contents of the information package are derived from a source material through digitisation or similar processes, then metadata about the source object SHOULD be included in the package. For example, if an information package contains a digitised image, then any metadata about the original analogue image should also be provided. This information is placed in the representation folder, under the metadata subfolder, in a folder named source. This may include technical metadata describing the image (resolution, color, etc.) or identifiers, provenance, and other information related to the source.

See also NBSIP11-19 in METS.xml (SIP)..
SHOULD0..1
NBSIPSTR18root_folder/schemas

The Information Package MUST have one subfolder at the root level named schemas, which MUST contain all XML schema documents for all structured metadata in the package. Schemas MUST NOT be placed in the representation folders.

This requirement is a stricter version of CSIPSTR15.
MUST1..1
NBSIPSTR19root_folder/documentation

If there is additional documentation necessary to make use of the contents of a package, it is recommended to include it. Such additional documentation SHOULD be placed in a folder at the root level named documentation. Examples of documentation include manuals for the system from which the data objects were exported, installation guides, and system usage requirements (hardware and software requirements). This additional documentation SHOULD be in a recommended file format, such as plain text (UTF-8) (e.g., XML, JSON, TXT) or PDF/A.

This requirement is a stricter version of CSIPSTR16.
SHOULD0..1
NBSIPSTR20The Information Package MUST only contain folders that are explicitly permitted by the NBSIPSTR requirements.

This requirement is a stricter version of CSIPSTR14.
MUST0..1
Last updated on - Github commit history ↗