Focus Group Minutes - 2001-aug-14

Title: 

Focus Group Meeting - 14 August 2001

Minutes - recorded by Joe 
Pato

 

Attendees

 

Joe Pato

Phillip Hallam-Baker

Don Flinn

Hal Lockhart

Prateek Mishra

Marlena Erdos

Carlisle Adams

Micah Lerner

Irving Reid

Tim Moses

Chris McLaren

Simon Godik

Helmut (VS)

Gil Pilz

Thomas Hardjono

Dave Orchard

 

Agenda

 

1)      administrivia

2)      current discussion items

3)      f2f #4 planning

 

1) Administrivia:

Don Flinn: will try to set aside 5 additional rooms for f2f #4 (initial 
10 already booked)

Prateek: When to freeze documents for F2F? Tuesday end of 
business.

Latest binding doc? - 04 (but plan to have a new 
revision next week)

 

F2F Agenda - items and duration

Dinner plans - joe put out e-vite on cuisine and interest

 

2) Discussion points:

 

Dave Orchard:

A)    Introduction of the object element as a container for resources and 
actions

a.      Prateek doesn't have issue; Phill would like to reorganize that section 
to make it maximally reusable by XACML.

b.      Prateek would like to keep the simple adjustments suggested by Dave for 
core-13 and then consider broader changes.

c.      Chris - intent is for the diagram to reflect the F2F; would rather get 
the document to represent the agreement, and to make broader changes at the next 
F2F (agreed)

B)     We should require an explicit authentication code and then address how to 
represent a query that represents <any>

a.      Phill views the agreement that the response needed to carry the code - 
not the query

b.      Dave - don't want to have an implicit wildcard want to have this 
explicit.

c.      Issue to be raised at the F2F <action: 
JOE>

C)    Min-occurs 

D)    Attribute Queries - attributes must have a value, min-occurs of 0 for 
value should not be allowed.

a.      Prateek - there is enthusiastic re-use of attribute element in both query 
and assertion. Suggestion is that the attribute query type should not refer to 
an unbounded number of attributes, possibly should refer to attribute name type 
(a single attribute name - not a collection; the query has a set of 0 or more 
attribute names) This would be coupled with making the cardinality of attribute 
type be changed to min-occurs 1.

b.      Dave - would prefer to have a distinguished value that would indicate the 
wildcard rather than to have an implicit value indicated by a min-occurs of 0. 
This to be an issue rather - <action: Hal will record as an 
issue>

c.      Phill would like to have agreement on the renaming changes before 
applying this change. Cf  Phill's 
message on renaming proposal.

d.      General agreement on the name changes - <action: Prateek will send 
e-mail with the proposed resolution in (b) above>.

 

 

Phill H-B

A)    Use 
of substitution groups - use of these by consumers of SAML will not be deterred 
by our avoidance. We should address the use of substitution groups by the SAML 
spec as an issue at the next F2F. 

a.      Prateek - request for assistance in "schematology" in XML. Looking for a 
separate evaluation of the aesthetics and practical (implementation) 
implications of the choices.

b.      Dave won't be at the F2F, but will comment on the 
results.

c.      While there isn't always enough experience, we need to make best 
attempt

 

Marlena

A)    working on a write-up on anonymity issues and on the configuration of the 
subject element

a.      issues around the use of the word "authentication"

b.      <action: marlena should be sending out a note 
shortly>

 

Prateek

B)     Issues with "signing" for the binding group

a.      SAML profile of DSIG - need to identify which kind of signature to be 
used, enveloping or not, attached or not, ... this is a blocking issue for now. 
Need a broader discussion of this issue.

b.      PHB - this needs to be part of the specification and part of  protocol probably - and good to 
coordinate with XKMS

C)    Box-caring of SAML request

a.      Will the core group address packaging multiple requests into a single 
object or is this resolved as a binding issue?

b.      Hal - what is the semantics, can there be cross component 
references

c.      Simon - concern that the absence of box-caring will lead to clear 
performance issues

d.      <action: PHB will write a note on how 1.0 might have some 
difficulties in expressing multiples - and possible directions to relieve this 
issue>.

 

3) F2F #4 Planning

 

 We will work on identifying items and time 
durations over the next week. <action: Joe will send out a 
solicitation for agenda items>

 

We expect to review active issues - but want people to establish clear 
priorities <action: Hal to make a first pass at prioritization of 
open issues and correlate to recent document revisions>

 

Some Items discussed during focus group are of clear interest for the 
face-to-face:

  

  
Address XML style issues: secure someone to lead us through both the 
  aesthetic niceties as well as the practical consequences of making certain 
  representation choices

  

  
DSIG use - develop a profile for how we should use XML 
  DSIG

  

  
Box-caring - consider how to deal with aggregation (if at all) of 
  messages - as a core issue, as a bindings issue

  

  
Address versioning considerations

  

  
Bindings and protocol issues - summary of current status. Prateek 
  requested that we address this topic early in the 
  face-to-face

  

  
Walk through of Schema and identification of new issues from the draft 
  specifications