Issue 577: Official NameSpaces of CRM Extensions?
Post by George Bruseker (15 December 20201)
I am wondering if anybody else struggles with what official namespace ot use for the CRM extensions. I'm not really sure how the situation stands. Should the minisites for each extension have a prominent place where they display the namespaces just so we all follow the same procedure? Do I miss what is already there?
Post by Nicola Carboni (20 December 2022)
The namespace to be used should be the xml:base value in the RDF document. Example:
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xml:lang="en" xml:base="http://www.cidoc-crm.org/cidoc-crm/CRMsci/">
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xml:base="http://www.ics.forth.gr/isl/CRMgeo/" xml:lang="en">
The confusion started because the namespace has changed over time
CRMdig 3.2.2 had
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xml:base="http://www.ics.forth.gr/isl/CRMext/CRMdig.rdfs/" xml:lang="en">
The latest version is
rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" xml:base="http://www.ics.forth.gr/isl/CRMdig/" xml:lang="en">
Generally they are both documented in prefix.cc, hence someone is still using the old ones.For clarifying the confusion, It is possible to write explicitly in the RDF itself the preferred namespace and prefix, using the properties vann:preferredNamespaceUri and vann:preferredNamespacePrefix . Example (in ttl) from VIR :
vann:preferredNamespacePrefix "vir" ;
vann:preferredNamespaceUri "http://w3id.org/vir#" ;
Post by George Bruseker (21 December 2022)
Thanks Nicola, that makes sense. I wonder if it is worth talking about what namespace the extensions have going forward. Taking CRMDig as an example. It arose from a project within which FORTH was a major partner and is an outcome of that work. It thus makes sense that it is registered under a FORTH namespace. But if it is considered an official extension, should it eventually have a namespace within the cidoc crm world for generally consistency / understandability / maintenance? May be worth a SIG conversation?
Post by Pavlos Fafalios (21 December 2022)
Dear George, all,
I agree that it is better to have namespaces under cidoc-crm.org for the official extensions, e.g.:
http://www.cidoc-crm.org/cidoc-crm/crmsci/ (or any other similar uri that starts with http://www.cidoc-crm.org/)
Also, these URIs, as well as the URIs of their classes and properties, should resolve to the latest "published version", based on the http request type (as we now do with the base model).
We discussed a bit about this on issue 460 (see point F):
When a new published version is available for one of the extensions (ideally aligned with crm 7.1.1), we can give it a try.
Post by Rob Sanderson (21 December 2021)
That seems like a big change, and long-term for the better, but disruptive in the shorter term while implementations change their namespaces.
A request, if we do go this route ... please don't nest namespaces, as it makes life much harder for processing.
For example, if CRM base is http://www.cidoc-crm.org/cidoc-crm/ then E55_Type would be http://www.cidoc-crm.org/cidoc-crm/E55_Type.
But then if sci is http://www.cidoc-crm.org/cidoc-crm/crm-sci/ processors need to be careful not to use the CRM prefix, with a term "crm-sci/O19_Encounter"
Thus, please, something like http://www.cidoc-crm.org/extensions/crm-sci/ as the namespace URI would be preferable.
In the 52nd joint meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9; 45th FRBR - CIDOC CRM Harmonization meeting; PF presented the state of the issue and the proposal to have the namespaces of compatible models under http://www.cidoc-crm.org/ (also avoid nesting namespaces). This will be implemented for published versions of extensions (assuming they have harmonized with CIDOC CRM v7.1.1). Namespaces for model extensions should look like that “http://www.cidoc-crm.org/extensions/crmsci” (for CRMsci f.i.)
Discussion: PR is not aware of potential conflicts with IFLA, they should know what their policy is before embedding LRM under this namespace. MR will attend a LIDATEC meeting in the beginning of March. She will raise the issue (namespace that LRMoo and FRBRoo will point to) and inform the SIG.
- The FR conceptual models fall under this: https://www.iflastandards.info/fr
- FRBRoo: https://www.iflastandards.info/fr/frbr/frbroo.html
- The space for the LRM models is here: https://www.iflastandards.info/lrm (So far it only features LRMer, but when ready, LRMoo will appear there with the prefix: https://www.iflastandards.info/lrm/lrmoo)
- Have all namespaces under http://www.cidoc-crm.org/ except for FRBR/LRMoo (discuss this once MR provides feedback). No hyphens in the model names.
- MR, PR to contact the team at FORTH (PF and ETz) inform them what the practice should be re. LRMoo and FRBRoo