Re: [apps-discuss] Comment on draft-ietf-appsawg-acct-uri-00.txt

Without the domain part, I would not know what server to go to to issue a
query.  What would a client do with "acct:paulej"?  If there is a "default
server" defined, then a default domain could equally be assigned.  The
latter usually is, anyway.

I think having input forms that accept just "paulej" are reasonable, but
client should construct fully-qualified URIs by adding "acct:" and
"@packetizer.com".

Paul

> -----Original Message-----
> From: apps-discuss-bounces@ietf.org [mailto:apps-discuss-bounces@ietf.org]
> On Behalf Of Peter Saint-Andre
> Sent: Tuesday, August 28, 2012 2:49 PM
> To: Graham Klyne
> Cc: apps-discuss@ietf.org
> Subject: Re: [apps-discuss] Comment on draft-ietf-appsawg-acct-uri-00.txt
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 8/28/12 9:41 AM, Graham Klyne wrote:
> > On 25/08/2012 02:21, Peter Saint-Andre wrote:
> >> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
> >>
> >> On 8/24/12 5:44 PM, Mike Jones wrote:
> >>> I agree that it's not a relative reference, in the RFC 3986 sense.
> >>> It's intended to be a context-dependent reference, as explained by
> >>> Phillip.
> >>
> >> Can you point to examples of context-dependent references in other
> >> URI schemes, or would the acct scheme be breaking new ground here?
> >
> > The one that springs to mind is file:, which (in my experience) is
> > almost always used in a localized context.  (Relative references are a
> > red herring here.) Another is about:, which is specifically defined
> > for accessing a local context.
> >
> > I think there's some conflation going on of my earlier comment that
> > you quoted ("URIs are intended to be a global namespace") with the
> > notion that every URI must be a global identifier with a unique
> > referent.
> >
> > I made my comment when speaking in support of keeping the domain name
> > part, this was on the basis that I could see acct: URIs likely to be
> > used in ways that escape their original context; e.g. as a hyperlink
> > in a document.  In the case of file: - if a file: URI (as opposed to a
> > relative reference, which is by definition
> > scheme-free) appears in (say) a web page, it's usually an error.
> >
> > That said, I will also concede that the web already lives with
> > contextualized URIs, so if there really is a compelling case for
> > domain-less acct: URIs, I could live with that.
> 
> I confess that I still don't understand the case for domain-less "acct"
> URIs. What exactly is the local context, and why would "acct"
> URIs be used in that context? For example, I could imagine that a local
> context might be accessing a device under my physical control or using a
> terminal to access infrastructure on a local network, but it's not clear
> to me why one would use an "acct" URI in those scenarios.
> What am I missing?
> 
> > I've said previously that I think the acct: scheme should be
> > associated with a specific dereferencing mechanism (i.e. WebFinger
> > - so it can be used as a generic hyperlink).  If domain-free acct:
> > URIs are allowed, then I think something should also be said about how
> > they should be dereferenced.
> 
> Agreed.
> 
> Peter
> 
> - --
> Peter Saint-Andre
> https://stpeter.im/
> 
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG/MacGPG2 v2.0.18 (Darwin)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
> 
> iEYEARECAAYFAlA9EqUACgkQNL8k5A2w/vymJwCgrBQliqt8GO6IGYw8na7ckabK
> BrEAnR/Vkj0AdiRO7heBqXuAOkba5YNa
> =+y5T
> -----END PGP SIGNATURE-----
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss