OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

xri message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: RE: [xri] XRI subject matching


Will makes a good point below about the fact that, under XRI 3.0, the two
http XRIs he lists as examples are http: URI bindings of the same XRI,
=willnorris.

And he's right that XRI processors will need to be able to do "XRI matching"
of different bound XRIs. To do that, they will need to three basic steps
(all of which will be in the Normalization and Comparison section of the XRI
3.0 Syntax spec I'm working on):

1) Unbind the bound XRI (i.e., remove the XRI from the base URI).
2) Normalize the XRIs to the same normal form (XRI-normal form, IRI-normal
form, or URI-normal form)
3) Compare the normalized XRIs.

However after John and I talked, I agree with John that we won't actually
need these XRI matching rules to be applied to XRD Subjects. When the
subject is a bound XRI, an XRD consumer should be able to treat it like any
other URI. We'll explain that in more detail in the XRI 3.0 Resolution spec,
which John and I will be working on as soon as the XRI 3.0 Syntax spec is
finished.

=Drummond 

> -----Original Message-----
> From: Will Norris [mailto:will@willnorris.com]
> Sent: Monday, August 24, 2009 2:18 PM
> To: XRI TC
> Subject: [xri] XRI subject matching
> 
> There has been a bit of discussion on the thread list XRD Subjects,
> most around the ongoing problem of what to do for host-meta
> documents.  To date, host-meta has been the only use-case for the
> match attribute on Subject, but I wanted to remind everyone of another
> -- XRI matching.  Drummond and John were supposed to talk about this a
> bit offline, so I'll let them respond with what they've come up with,
> but very preliminary discussions I had with Drummond indicated that
> the match attribute would be very useful for XRI matching.  I'll be
> the first to admit I don't fully grok XRI syntax, much less the latest
> 3.0 draft, so someone can explain more if necessary.  All I do know
> however, is that the following two URIs represent the same
> "=willnorris" XRI:
> 
>    http://xri.net/=willnorris
>    http://xri.fullxri.com/=willnorris
> 
> There is no way to know that these should be considered equivalent
> without some special processing rules.  They are certainly not
> equivalent strings, and a "beginswith" comparison doesn't work
> either.  This means that XRI Resolution 3.0 will need to define a new
> value for the match attribute that means, "This thing is an XRI.
> Compare it against other URIs using the following rules...".  If we
> drop the match attribute in favor of something like <Host> for host-
> meta, then we're leaving XRI resolution out in the cold.  While I
> believe that "beginswith" Subject matching for host-meta is the worst
> possible solution (except of course for all the others), I want to
> make sure folks remember that we do have at least one other use case
> to keep in mind.
> 
> -will
> 
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  Follow this link to all your TCs in OASIS at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php




[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]