Re: [cmis] [OASIS Issue Tracker] Created: (CMIS-39) Remove propertyid of the spec and have only name as unique identification of a property

[Julian Reschke <julian.reschke@greenbytes.de>]

OASIS Issues Tracker wrote:
> Remove property id of the spec and have only name as unique identification of a property
> ----------------------------------------------------------------------------------------
> 
>                  Key: CMIS-39
>                  URL: http://tools.oasis-open.org/issues/browse/CMIS-39
>              Project: OASIS Content Management Interoperability Services TC
>           Issue Type: Improvement
>             Reporter: Jens Huebel
> 
> 
> The spec mentions two different concepts to uniquely identify a property. It has a notion of a name and an id. The id is exposed as part of the getProperties() method of the object service but it seems never to be used as input of any method (e.g. updareProperties). So what is the difference between name and id and do we really need both concepts?

This reminds me of another thing I've been asking myself recently...: 
most metadata models I know (WebDAV, JCR, Dublin Core, RDF) use either 
Expanded Names (namespace + local name) or URIs as identifier.

Just having a simple string will make it *very* hard to interoperate 
with these systems (for instance, how to expose RDF/XMP metadata in 
CMIS), and will kill distributed extensibility.

Feedback appreciated,

Julian