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


Help: OASIS Mailing Lists Help | MarkMail Help

xdi message

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

Subject: Minutes:Joint XRI & XDI TC Telecon 10AM PT Thursday 2007-10-25

Following are the minutes for a joint unofficial telecon of the XRI and XDI
TCs at:

Date:  Thursday, 25 October 2007 USA
Time:  10:00AM - 12:00PM Pacific Time

Event Description:
Weekly unofficial joint call of the XRI and XDI Technical Committees.


Gabe Wachob
Wil Tan
Markus Sabadello
Drummond Reed
John Bradley



Drummond gave a brief update on his spec review session yesterday with
OASIS staff member Mary McRae, who is responsible for all TC spec
preparation. Mary said the spec looked good, the Conformance section met
OASIS requirements, and we went through the proper formatting and
requirements for all the appendicies. So we are ready for a Committee Draft
vote as soon as we finish all revisions.

# DRUMMOND to update all TODOS in the spec that were awaiting Mary's input.


We closed on the following seven issues that have been raised in feedback
about the ED06 spec posted at:

***** #1: Encoding of + signs and spaces in HXRIs.

* We will specify that HXRIs MUST be encoded using the following steps in

	1) Encode the HXRI, including all query parameters, in URI-normal
form (which requires percent-encoding of invalid URI characters), 
	2) Percent encode all percent signs, and then 
	3) Percent-encode any ampersand or semicolon inside HXRI query

* We will also specify that proxy resolvers MUST ensure that exactly one
pass of percent unencoding is performed on the HXRI after receipt. This
means that if percent unencoding not automatically done by the Web server
hosting the proxy resolver, the proxy resolver must do it.

* Because any space that was originally percent-encoded as %20 will be
encoded as %2620 in step 2 above, then unencoded back into %20, this means
that any SPACE that appears in the HXRI when received by the proxy resolver
code represents a + sign (which was turned into a space by the Web server. 

***** #2: Leading slash in <Path> element.

* Our conclusion was to change ED06 so that the leading slash is explicitly
required in the Path element, and also to specify that if it is not present,
one MUST be added before comparison.

* We also decided that the XRI Resolution 2.0 spec will treat an XRI
authority with and without the trailing slash as equivalent, i.e., @example
and @example/ are equivalent for purposes of comparison.

* This also means that both no trailing slash and a single trailing slash
will be equivalent to a null path, and will match ANY of the following three
Path element instances:

<Path match="null" />

***** #3: Status element and SAML signature conflict issue.

* The conclusion is to adopt the "sidestep" approach per the mailing list
discussion, i.e., the ED06 spec will remain the same except we will add an
explanation that consuming applications that wish to independently verify
the SAML signature can either deconstruct the XRDS file OR request the XRD
directly from the SAML authority server.

***** #4: Treatment of empty SEP selection elements that have no match
attribute. ED06 section 12.6.4 says they are treated as if match="default".
Wil proposes they be treated as if match="null".

* There was consensus that Wil's proposal is correct. Drummond will make the

***** #5: Treatment of ALL attributes with empty/null string as an error.
ED06 says they are treated as being absent. Wil proposes this be treated as
an explicit error.

* There was consensus that the behavior for empty elements should be
dictated by the rules of W3C XML Schema, and thus will typically produce a
malformed XRD error. We will say this in ED07 (and change the numerous
places where we currently say that an empty attribute is equivalent to the
attribute being absent). We will also say that empty element values is NOT

***** #6: ED06 section 12.7.6 line 2497: Should we allow all values of the
append=attribute when constructing an authority resolution service endpoint

* The consensus was that to keep things simple, no special rule will be
imposed, however we will add a warning to XRD authors that, in an authority
resolution service endpoint URI, using any value for the append attribute
except append="authority" is NOT RECOMMENDED.

***** #7: Clarification of Ref and CanonicalID verification rules.

* John pointed out that all Refs processed BEFORE the completion of
authority resolution do not change the CanonicalID verification chain, while
any Ref processed AFTER the completion of authority resolution will start a
new CanonicalID verification chain.

* We agreed to add text to the Redirect and Ref section to point this out,
and also point out the greater flexibility of using a Ref in a SEP, inluding
the ability to put it in a default SEP.

# DRUMMOND to roll all these changes into ED07.


We reviewed Gabe's proposed RelaxNG schemas:


We made the following decisions regarding element order:

* ServerStatus will go AFTER Status
* ( Redirect | Ref ) will go AFTER all synonym elements and BEFORE Service

The other changes that need to be made are:

* To XRDS schema: add redirect attribute.
* To XRD schema: add support for other attributes from other namespaces.

We then discussed what steps Gabe recommends to validate his RelaxNG

	1) Run it through a validator like JING

	2) Validate all examples in the spec using this validator.

	3) Validate XRDs in the wild against it.

# GABE and WIL to work together on validation of the schema, with the goal
of having it completed and ready to copy into ED07 by next Monday night.


We need to make sure there are not any IPR-related issue about the Yadis 1.0
spec being incorporated as Section 6.

# DRUMMOND to work on this.


Mary McRae explained that each TC sets its own policy for this.

After discussion, we agreed on the following layout:

	A) Contributors: list all active contributors to the spec besides
the editors.
	B) TC Members: list all current TC members except any TC member that
asks not to be listed.
	C) Other Acknowledgements: List everyone currently acknowledged in
the 2.0 cycle, plus anyone else who has contributed but is not a TC member.

# DRUMMOND to send an email to the list to provide notice about this


The schedule is:

* Gabe and Wil to complete RelaxNG schema testing and send Drummond the
revised RelaxNG schemas by Monday night.
* Drummond to issue ED07 on Tuesday with all changes reflected.
* Next Thursday's TC call will be to close on ANY AND ALL final feedback on
* Drummond will publish WD11 Final on Friday.
* We will begin the Committee Draft vote on Friday (it runs for one week).


To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  You may a link to this group and all your TCs in OASIS

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