RE: [xri] RE: Single delegation character

["Dave McAlpin" <Dave.McAlpin@epok.net>]

I guess I agree, although I'm not sure about "where control lies". All
the XRI spec has to say about @idcommons/fen is that @ will return an
XRID for idcommons, and that XRID will in turn describe local access
protocols for the remainder of the path. It's perfectly fine for
idcommons to expose exactly one local access protocol that asks for the
next segment and promises to return an XRID.

That's actually one of the reasons we reserved dot through the path
portion of XRIs, so that local access could continue resolving the path
one chunk at a time if it chose to. Imagine a 1.0 style xri in the form

xri:@a.b.c/d.e.f/g.h

The XRI spec says you'll publicly resolve @a.b.c and end up with local
access descriptors for the remaining path. In other words, like DNS
based URIs, the authority portion is considered public and the path
portion is considered private. But what if you wanted to continue that
pattern, and make portions of the path more "public" than others? What
if, for example, you wanted to "resolve" d.e.f without giving up
information about g.h? It would be fine for c to publish a local access
service defined exactly like standard resolution. At that point you'd
have something functionally equivalent to xri:(xri:@a.b.c).d.e.f/g.h,
where the network endpoint for the root (xri:@a.b.c) was known from the
local access descriptor returned by the previous round of resolution. 

Sorry for the alphabet soup there. The point is that up to the first
slash, whatever the replacement for dot turns out to be has defined
semantics related to resolution. After the first slash, slash and the
replacement for dot are reserved but their meaning is defined only in
terms of a particular local access protocol.

Dave

> -----Original Message-----
> From: Victor Grey [mailto:victor@customdynamic.net]
> Sent: Friday, June 04, 2004 4:06 PM
> To: Dave McAlpin
> Cc: Fen Labalme; Loren West; xri@lists.oasis-open.org
> Subject: Re: [xri] RE: Single delegation character
> 
> Dave McAlpin wrote:
> > Not really. @idcommons/fen just says resolution turns into local
access
> > at idcommons. There's no reason that local access can't say "give me
> > the
> > next segment and I'll return an XRID".
> 
> But isn't it all about being able to express where the control lies? I
> thought the point was that we could say "resolution turns into local
> access here" OR "resolution references a different authority here".
> 
> BTW - among not only potential users, but developers as well (who are
> imo the real audience for what we're creating), my anecdotal results
> are that ! is universally hated (" ! indicates 'alert - something
> wrong' -- bad feng shui" was one comment). We will try to do a
slightly
> more rigorous survey this weekend with a few hundred people - many of
> them developers.
> 
> -Victor
>