Skip to content

Conceptual Description of the Level 1 C2M2

authors: Rick Wagner

maintainers: Rick Wagner

version: initial draft

license: CC0 1.0 Universal (CC0 1.0) Public Domain Dedication


Resources (files, subjects, biosamples, etc.) are uniquely named in the context of the C2M2 using a namespace and an local name (the id column of the entity tables). When combined, the namespace and local name form a unique name for the resource through the CFDE. In principle, this could be done using a only a full URI for the resource. In practice, the use of namespaces is a mechanism to allow DCCs to name their resources in a simple manner without collision. The CFDE CC defines a process for DCCs to define their own namespaces in a manner that avoids collions, without requiring CFDE CC involvement. DCCs are welcome to engage the CFDE CC with questions before investing significant effort in creating their ETL process.


The namespace must be a valid URI (not necessarily resolvable) that is under the control of the DCC or a trusted naming authority, and may contain an intial path information. Examples of valid namespaces * * *,2020:/ *,2020:samples#

The namespace itself does not need (nor should) have any semantic meaning on its own, though the namespaces table in the C2M2 may contain descriptive text. Nor does the namespace define relationships between resources, such as project membership or type. Files, biosamples, subjects, and collections declare what project they are associated with. The relationship between entities, such as a file produced from a biosample, is defined separately from what project or namespace those entities are associated with.


The local name of a resource is concatenated with the namespace to form the full resource name. To ensure this, the namespace must be a valid URI and may include a trailing separation character, such as ? or #. The local name of a resource may have any syntax, so long as when concatenate with the namespace the full names is also a valid URI.

Given the local name 8675/REAMDE and the previous namespace examples, the full resouces names would be: * * *,2020:/8675/REAMDE *,2020:samples#8675/REAMDE


There are tools for validating and introspecting URIs, like the Python rfc3986 package. Using our examples above, we can valid the namespaces, the full names, and look at the breakdown of the elements.

import rfc3986

URIReference(scheme='https', authority='', path='/', query=None, fragment=None)

URIReference(scheme='https', authority='', path='/samples/', query=None, fragment=None)

URIReference(scheme='tag', authority=None, path=',2020:/', query=None, fragment=None)

URIReference(scheme='tag', authority=None, path=',2020:samples', query=None, fragment='')

URIReference(scheme='https', authority='', path='/8675/REAMDE', query=None, fragment=None)

URIReference(scheme='https', authority='', path='/samples/8675/REAMDE', query=None, fragment=None)

URIReference(scheme='tag', authority=None, path=',2020:/8675/REAMDE', query=None, fragment=None)

URIReference(scheme='tag', authority=None, path=',2020:samples', query=None, fragment='8675/REAMDE')

Namespace Selection

There are two recommended ways to choose a namespace, tag URIs or HTTPS and an FQDN. Tag URIs and HTTPS URLs using an FQDN allow DCCs to define their own namespaces, while splitting a persitent identifier permits enternal use of the full resource name.

Tag URIs have a simple syntax with the scheme "tag:", the entity defining the URI, and the local name: tag:<naming authority>:<local name> The <naming authority> may be an FQDN and a date that FQDN was managed by the DCC. E.g., for's Project A, their namespace may be,2020:

HTTPS URLs are similar to XML namespace, where the DCC must have control of the FQDN, but the URI does not need to be a resolvable URL. could use this to declare a namespace for their samples from the Project A study.


Namespaces are declared in the id_namespace table

Column Name Type Description
id URI A globally unique ID representing this namespace. Used as a foreign key by entities.
name slug A short, human-readable, machine-read-friendly label for this namespace.
description string A human-readable description of this namespace.


  "id":          ",2020:samples",
  "name":        "Samples from the Project A study",
  "description": "The sample identifiers sample were derived from the
        Project A accession numbers."

The descriptive text (name and description) should related to the namespce and the local names of the resources.


This section provides a concise overview of the mechanism to declaring a namespace associated with a DCC in the context of the CFDE project. It is an important harmonization process, which aims to avoid collisions between identifiers, while enhancing interoperability and findability. It also represents a key process in mapping data into the C2M2 model, thereby getting ready for a full ETL process from a DCC to the CFDE model and future indexing in the CDFE Deriva system.