RE: identifier or dc:URI?

 

Hi,

> > IMO if we have a clear use case for non-URI identifiers 
> then we should 
> > choose "identifier", if not we should choose "URI"
> 
> For earl:WebContent it is clear that we need a URI to 
> describe the resource. However, I can imagine some use cases 
> for dc:identifier too. For example a hash or checksum of the 
> resource to compare it at a later fetch of the same URI.
> 
> As to earl:Software, it seems to be the opposite approach: an 
> identifier such as the build or version number may be the 
> best identifier, and URI may sometimes be available too (as a 
> location to download or find more information about the software).

The only problem I see is that identifiers are supposed to be
unambiguous, and a build or version number (or a checksum) is not
because it could be useful to identify a piece of Software if you know
other information (e.g. name, developer, etc.), but not by itself. For
example:

- WatchFire Bobby v5.2 can identify specific software
- Microsoft IExplorer v5.2 can also

But v5.2 can not (Bobby or IExplorer?)

Additionally the fact of having several identifiers for something could
be a bad idea because people are used to unique identifiers (one per
thing)


Regards,

CI.

Received on Wednesday, 21 June 2006 13:14:12 UTC