Re: [p2-dev] URLs, URIs, and IDs (oh my)

[Scott Lewis <slewis@xxxxxxxxxxxxx>]

Hi Doug,

Schaefer, Doug wrote:
> Ah, you are right. I stand corrected.
>
> >From the resource standpoint, and EFS in particular, the URI works
> great. The extensibility comes by adding layers on top of the URI based
> on the scheme, in our case by mapping scheme to IFileSystem which
> manages the additional functionality. Clients obviously need to know
> about the extensions to call them. But I think we're in control of that.
>   

One issue here is the use of java.net.URI rather than emf.URI (no 
interface/impl independence).  But yes, I agree URI for a remote file 
system/file identifiers makes complete sense.

> Anyway, just an example of where URIs in action. I'm not familiar enough
> with ECF to know whether this applies there, though.
>   

Yes, I think it does.  We have (for example) a file transfer provider 
that uses EFS for implementing the asynchronous transfer (using Jobs).  
The ECF IFileID for this provider/Namespace is:

efs:http://foo.bar/path/to/resource.ext

or

efs:<efs filestore uri>

Of course this can/could also be represented as a URI.

Scott