RE: Subject of a host-meta XRD (was: <SubjectTemplate>)

[Eran Hammer-Lahav <eran@hueniverse.com>]

I am looking for feedback from the TC about the questions below as well as gauging the TC's appetite for venturing into URI sets and rules (either via a URI template of some sort or a more structured XML schema similar to POWDER).

So our options are:

1. Define a new element(s) in XRD to allow a resource set as a subject. This solves all the semantic issues because it eliminates the concept of a host. The XRD in that case applies to any resource that matches its rules. We can offer a very limited URI sets functionality that will be useful for many simple use cases. Something like:

<XRD>
	<SubjectSet>
		<Authority>example.com:80</Authority>
		<Schemes>http https</Schemes>
		<Suffix>*</Suffix>
	</SubjectSet>
</XRD>

In which the authority is used to sign (wildcards are allowed the same way they are allowed as certificate subjects), schemes are space delimited, and suffix a regex (defaults to all omitted). This makes it easy to take any URI and match it against the rules:

http://example.com/resource/1 --> scheme: http, authority: example.com:80, suffix: /resource/1

2. Define a new URI for the host (as a non network resource) and use it as the subject of the host-meta XRD. LRDD (where this will be defined) will make it clear that anything other than links with URI templates and a describedby relation (alone or in addition to others) are left undefined. The problem is, while it was easy to do that using Link-Pattern, it is not so easy with <Link>. Also, while LRDD can register the new URI scheme, it will not go without resistance.

3. Use a different schema for host-meta that is compatible with XRD for the purpose of signatures. Something like:

<Host-Meta>
	<Host>example.com:80</Host>
	<!-- Signature stuff here same as XRD -->
	<Link-Pattern rel="describedby" pattern="http://example.com/meta?uri={%uri}"; type="application/xrd+xml" />
</Host-Meta>

My vote is for #1 and then #3. While #3 is easier, #1 offers a more consistent framework and allows future uses for host-meta since it is a fully featured XRD.

EHL



> -----Original Message-----
> From: Drummond Reed [mailto:drummond.reed@cordance.net]
> Sent: Wednesday, July 01, 2009 8:37 AM
> To: Eran Hammer-Lahav; 'XRI TC'
> Subject: RE: Subject of a host-meta XRD (was: <SubjectTemplate>)
> 
> Agreed on all points.
> 
> So what's the gameplan for coming to closure on this issue? Should I
> put it
> on the agenda for tomorrow's telecon?
> 
> =Drummond
> 
> > -----Original Message-----
> > From: Eran Hammer-Lahav [mailto:eran@hueniverse.com]
> > Sent: Tuesday, June 30, 2009 11:00 PM
> > To: Drummond Reed; 'XRI TC'
> > Subject: RE: Subject of a host-meta XRD (was: <SubjectTemplate>)
> >
> > The issue is still open, but I am not sure we have nails the exact
> problem
> > yet.
> >
> > If you recall about 6 months ago we had a discussion about whether
> Links
> > in host-meta apply to individual resources belonging to the host
> under
> > LRDD. According to the host-meta spec, they do, but LRDD I explicitly
> > chose to ignore them for the purpose of looking for descriptors. A
> few
> > days ago Dirk raised the same questions.
> >
> > We had all these issues before with the text-based host-meta but they
> were
> > just not that obvious because we were unable to apply all the
> thinking
> > that went into XRD there. But now that we reformatted host-meta to
> use
> > XRD, it all surfaced.
> >
> > The two main questions are:
> >
> > 1. What is a 'host' (or 'site' or 'authority' etc.)?
> > 2. How do we identify it?
> >
> > So far we defined host as either:
> >
> > 1. Any resource belonging to the combination of domain/port/protocol
> (i.e.
> > a set), or
> > 2. An abstract concept of authority restricted by domain/port and
> possibly
> > protocol
> >
> > A Link to one has a very different meaning than the other. In fact, I
> > don't know what a Link means for #2...
> >
> > If we decide a host is a 'set', we should find a way to describe a
> set of
> > URIs (which will be useful for other purposes). If we decide a host
> is a
> > resource (of some sort), we should find a URI to identify it.
> >
> > I don't think there is right or wrong here. It is just a question of
> which
> > option is easier/more productive/reusable.
> >
> > EHL
> >
> >
> > > -----Original Message-----
> > > From: Drummond Reed [mailto:drummond.reed@cordance.net]
> > > Sent: Tuesday, June 30, 2009 10:47 PM
> > > To: Eran Hammer-Lahav; 'XRI TC'
> > > Subject: Subject of a host-meta XRD (was: <SubjectTemplate>)
> > >
> > > Eran,
> > >
> > > I knew as soon as the question "What's the subject of a host-meta
> XRD?"
> > > came
> > > up that we were headed into W3C httpRange-14 territory. For those
> who
> > > have
> > > never heard about this "Web identity crisis", see:
> > >
> > > 	http://www.w3.org/2001/tag/doc/httpRange-14/2007-05-31/HttpRange-
> > > 14
> > > 	http://norman.walsh.name/2005/06/19/httpRange-14
> > > 	http://www.garshol.priv.no/blog/125.html
> > >
> > > There are dozens more references where those came from, as I'm sure
> > > those
> > > who have stepped into this quicksand know. Just Google "httpRange-
> 14".
> > >
> > > So I'm not going to try to give any philosophical answers here,
> only
> > > practical ones. On that basis, my observations:
> > >
> > > 1) Both the <XRD:SubjectTemplate> suggestion (see below) and the
> Powder
> > > approach (essentially another way of desribing a URI template using
> > > individual XML elements for each constraint) seem like reasonable
> > > options. I
> > > prefer SubjectTemplate because it's less complex. But I'm not sure
> what
> > > the
> > > SubjectTemplate value would be that describes "the authority for
> this
> > > domain" vs. any specific resource in that domain. Would it just be
> a
> > > wildcard?
> > >
> > > 2) A second option is to use either a fragment, or an empty
> fragment. I
> > > prefer the latter for this particular use case. In other words, if
> > > http://example.com identifies a specific information resource
> (e.g.,
> > > the 200
> > > response you get back), then http://example.com# could describe the
> > > abstract
> > > concept of http://example.com (a non-information resource).
> > >
> > > 2) A third option -- mentioned frequently in the httpRange-14
> debate --
> > > is
> > > using another URI scheme intended exclusively to represent non-
> > > information
> > > resources. (Hmmmm, where could we find something like that? ;-) Of
> > > course,
> > > it's ironic that due to W3C TAG's input, XRI is no longer actually
> > > another
> > > URI scheme, but an identifier syntax that is bound to a base URI to
> > > produce
> > > a valid URI. I posted earlier about what the bound http: XRI that
> > > described
> > > "the current authority" would look like: http://xri.net/$.
> > >
> > > In any case, all three of these options would appear to work. Have
> you
> > > decided on one? Is the issue still open?
> > >
> > > =Drummond
> > >
> > > > -----Original Message-----
> > > > From: Eran Hammer-Lahav [mailto:eran@hueniverse.com]
> > > > Sent: Sunday, June 28, 2009 12:37 PM
> > > > To: XRI TC
> > > > Subject: [xri] <SubjectTemplate>
> > > >
> > > > This idea isn't new but given the need to solve the host-meta
> Subject
> > > use
> > > > case, I would like to know what others here think about it.
> > > >
> > > > EHL
> > > >
> > > > -----Original Message-----
> > > > From: uri-request@w3.org [mailto:uri-request@w3.org] On Behalf Of
> > > Eran
> > > > Hammer-Lahav
> > > > Sent: Sunday, June 28, 2009 12:32 PM
> > > > To: Erik Wilde; uri@w3.org
> > > > Subject: RE: URI for abstract concepts (domain, host, origin,
> site,
> > > etc.)
> > > >
> > > > Using a URI template is one option being considered (XRD already
> has
> > > a
> > > > <URITemplate> element under <Link> so the syntax is already part
> of
> > > XRD).
> > > > However, that requires either creating a new element (like
> > > > <SubjectTemplate>) or changing the XML schema type for <Subject>
> > > which
> > > > currently does not allow anything but valid URIs.
> > > >
> > > > But before we consider that, I wanted to see if there was an easy
> > > solution
> > > > for describing such resources with a URI.
> > > >
> > > > EHL
> > > >
> > > > > -----Original Message-----
> > > > > From: uri-request@w3.org [mailto:uri-request@w3.org] On Behalf
> Of
> > > Erik
> > > > > Wilde
> > > > > Sent: Sunday, June 28, 2009 11:43 AM
> > > > > To: uri@w3.org
> > > > > Subject: Re: URI for abstract concepts (domain, host, origin,
> site,
> > > > > etc.)
> > > > >
> > > > > hello.
> > > > >
> > > > > Eran Hammer-Lahav wrote:
> > > > > > Let me try explaining my use case again, this time without
> any
> > > > > overloaded terminology or proposed solutions.
> > > > > > XRD is a document format for describing resources. It looks
> like
> > > > > this:
> > > > > > <XRD>
> > > > > > 	<Subject>http://example.com</Subject>
> > > > > > 	<Type>http://example.org/type/blog</Type>
> > > > > > 	<Link>
> > > > > > 		<Rel>author</Rel>
> > > > > > 		<URI>http://example.com/author</URI>
> > > > > > 	</URI>
> > > > > > </XRD>
> > > > > > Without getting too much into XRD, this short descriptor
> > > describes
> > > > > the resource identified by 'http://example.com'. It includes
> one
> > > type
> > > > > indicator (a made up example meant to mean this resource is a
> > > blog),
> > > > > and one link to the author's page.
> > > > > > I want to use this document format to describe rules that
> apply
> > > to
> > > > > all resources which belong to an HTTP host (as defined by 2616:
> a
> > > > > domain/address and port combination). The problem is, <Subject>
> > > > > requires a URI and currently there is no way to identify this
> set
> > > of
> > > > > resources (http://domain:port/*) as a valid URI.
> > > > > > What I don't want to do is use an exception such as 'if the
> URI
> > > > > begins with X, treat it as a rule and not a valid URI'...
> > > > >
> > > > > given this new description, isn't what you're looking for a URI
> > > > > template
> > > > > language for XRD? maybe not exactly the one currently proposed
> by
> > > > > http://tools.ietf.org/html/draft-gregorio-uritemplate-03, but
> isn't
> > > > > that
> > > > > close to what you want? a template notation would also nicely
> > > address
> > > > > the case mentioned already where the host scope would be too
> > > general.
> > > > > but then again, a URI template is not a URI, so you could use
> it in
> > > the
> > > > > context of XRD, but not as a standalone URI....
> > > > >
> > > > > cheers,
> > > > >
> > > > > erik wilde   tel:+1-510-6432253 - fax:+1-510-6425814
> > > > >         dret@berkeley.edu  -  http://dret.net/netdret
> > > > >         UC Berkeley - School of Information (ISchool)
> > > >
> > > >
> > > >
> > > > -----------------------------------------------------------------
> ----
> > > > 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
> > >
>