Meeting started at 8:04 PST
====================================================================
Roll Call
Voting Members:
---------------------------------
Alejandro Abdelnur Sun yes
Sasha Aickin Plumtree yes
Subbu Allamaraju BEA yes
Olin Atkinson Novell yes
Atul Batra Sun yes
Amir Blich SAP yes
Chris Braun Novell yes
Rex Brooks Starbourne yes
T.J. Cox Novell yes
William Cox BEA yes
Brian Dirking Stellent no
Michael Freedman Oracle yes
Ross Fubini Plumtree yes
Richard Jacob IBM yes
Jon Klein Reed-Elsevier yes
Andre Kramer Citrix yes
Alan Kropp Vignette Corporation yes
Carsten Leue IBM yes
Dan Machak Tibco yes
Madoka Mitsuoka Fujitsu no
Adam Nolan Reed-Elsevier yes
Petr Palas Moravia IT yes
Sunit Randhawa Fujitsu no
Nigel Ratcliffe Factiva yes
Thomas Schaeck IBM yes
Gennady Shumaker SAP yes
Yossi Tamari SAP Portals yes
Gil Tayar WebCollage no
Rich Thompson IBM yes
Charles Wiecha IBM yes
Total voting members: 30
Voting members in attendance: 26 (87%)
A quorum was present.
Members on Leave Of Absence
----------------------------
Raj Ramesh CommerceOne LOA
Joe Rudnicki U.S. Navy LOA
Steven Smith Capitol College LOA
Eric van Lydegraf Kinzan LOA
====================================================================
The corrected minutes from 3/04/03 were accepted
====================================================================
Summary of Change Requests
202 - Provide reference to the schema datatype defs ACCEPTED
203 - Wording for resend cached data ACCEPTED
211 - NamespacePrefix is now a separate field ACCEPTED
212 - markupCharacterSet is now an array ACCEPTED
217 - Update section reference ACCEPTED
219 - Producer invalidation of registrationHandles ACCEPTED
225 - Specifying preferred character set ACCEPTED
226 - Resource URL ACCEPTED
227 - Explicitly specify the relevant portlet url params ACCEPTED
187 - Cacheability and perform*Interaction ACCEPTED
195 - Add registrationMethods NOT ACCEPTED
173 - What does it mean to support registration? ACCEPTED
162 - Require Consumers to support UTF-8? NOT ACCEPTED
143 - Properly encode '&' in examples and BNF
ACCEPTED
177 - Templates MUST send placeholders ACCEPTED
164 - Namespacing Portlet URL Parameters ACCEPTED
138 - How does info get to proxied resources ACCEPTED*
193 - Asserting multiple standard user categories ACCEPTED
192 - Prevent localization of standard user cat. names NOT ACCEPTED
200 - Clarify response to ignoring security request NOT ACCEPTED
201 - PortletDescription applies to clones ACCEPTED
204 - Fields for cache key NOT ACCEPTED
141 - Add previous windowState and mode? Deferred (JSR-168)
205 - Template requirements NOT ACCEPTED
206 - Values for extensible items such as modes ACCEPTED
All Change Requests related to Conformance Statements ACCEPTED
* partially accepted, partially deferred - see detail below
====================================================================
Summary of Action Items:
-----------------------
- Andre will provide a proposal for getting a resource
via a SOAP call. (See CR 138)
- Alan Kropp will consider helping out on writing the primer.
(See CR 204)
- Rich will release a new version of the specification by
next week.
====================================================================
Small editorial/grammar changes
202 - Provide reference to the schema datatype definitions
Section 5.1.15 / Page 24 / Line 6 / Andre Kramer
203 - Wording for resend cached data
Section 6.1.1 / Page 26 / Line 36 / Andre Kramer
States preference for the wording from initCookie() fault
processing
211 - NamespacePrefix is now a separate field
Section 5.1.11 / Page 20 / Line 44 / Rich Thompson
Reference to the additional field is now needed in this
conformance statement
212 - markupCharacterSet is now an array
Section 6.1.9 / Page 30 / Line 30 / Rich Thompson
Grammar didn't get updated when the field became an array.
217 - Update section reference
Section 6.4 / Page 41 / Line 18 / Rich Thompson
Current reference is wrong as things have moved around
219 - Producer invalidation of registrationHandles
Section 7.2 / Page 48 / Line 36 / Rich Thompson
Rephrases for clarity
225 - Specifying preferred character set
Section 10.1 / Page 55 / Line 16 / Rich Thompson
Correct optionality & the change to array of char sets.
226 - Resource URL
Section 10.2.1.1.4.1 / Page 58 / Line 31 / Rich Thompson
Make the conformance stand-alone
227 - Explicitly specify the relevant portlet url parameters
Section 10.2.2.11 / Page 62 / Line 10 / Rich Thompson
Be explicit about what portlet URL parameters are being
referred to.
Resolution: Accepted
====================================================================
187 - Cacheability and perform*Interaction
Section 6.3.x / Page 39 / Line 9
Two alternatives proposed for clarifying how the protocol has
clues for interactions impacting caching. Discussion has centered
around whether we want anything in the spec until we define full
invalidation caching.
Resolution: Accepted - Will insert a statement that Consumers must
always propatage end-user actions to the Producer.
====================================================================
195 - Add registrationMethods
Section 5.1.18 / Page 23 / Line 7
Proposal adds a field to enable a Producer to provide guidance
for how to register. In-band technique is identified by the portType
name as this eliminates dependencies on particular bindings while
not inventing yet another name.
Resolution: Not Accepted - discovery of the registration method
would be in same domain as discovery of the Producer
Service.
====================================================================
173 - What does it mean to support registration?
Section 8.3 / Page 52 / Line 20-22
We have a conformance statement regarding Producers supporting
registration. It currently points at the requireRegistration
flag ... not an exact match.
Resolution: Accepted - Will rework paragraph: "if consumer has
registered then it MUST ...."
====================================================================
162 - Require Consumers to support UTF-8?
Section 10.1 / Page 55 / Line 22
Proposal notes that we imply the answer to this question is yes
and suggests adding a conformance statement to that effect.
Resolution: Not Accepted. However, whereever UTF-8 is mentioned, it
will be replaced with UTF-8 or UTF-16
====================================================================
143 - Properly encode '&' in examples and BNF
Section 10.2.1 & 10.2.2 / Page 58 & 61 / Line 15 & 24
Question raised is whether or not the fragment containing a URL
destined for consumer rewriting needs to be strictly XML compliant.
If so, our '&' should be changed to '&'.
Resolution: Accepted - portlets have the responsibility to encode
the results of a template to be compatible with the
markup type they are generating. An example of this is
the encoding of entity refs within text/xml markup.
(This will be in the form of a conformance statement)
====================================================================
177 - Templates MUST send placeholders
Section 10.2.2.1 - 10.2.2.8 / Page 61
Change 'SHOULD' to 'MUST' for the placement of fundamental
parameters in action, etc templates (not for the default templates).
Resolution: Accepted
====================================================================
164 - Namespacing Portlet URL Parameters
Section 10.3 / Page 64 / Line 18
Alternate text suggested for clarity reasons.
Resolution: Accepted - Rich will reword and modify conformance
language
====================================================================
138 - How does info get to proxied resources
(Insert a new section 10.4?)
Section 10.4 (insert a new section at this point in the spec)
In addition to a new section to discuss proxied resources, this
request raises the question of how does a resource get access to
information. Current discussion has noted three existing techniques
and proposed one new technique. The three existing are:
1) Cookies that are properly defined/processed relative to cookie
domains.
2) URL-encoding of the information.
3) URL-encoding of an indirection to the information.
The proposed new technique is to define HTTP headers that carry
UserContext and UserProfile information. The portlet could use
portlet URL parameters to control whether or not these two
items are supplied to a resource.
Resolution: 1) Accepted - When a Consumer claims that it supports
cookies it must also support cookies for proxied
resources.
2) Not Accepted
3) Deferred - Andre will provide a proposal for getting
a resource via a SOAP call.
====================================================================
193 - Asserting multiple standard user categories
Section B.2 / Page 77 / Line 25
Adds explicit statements that multiple std user categories can be
asserted and that there are no correlations between them.
Resolution: Accepted - Will remove the second sentance in B.2
====================================================================
192 - Prevent localization of standard user category names
Section 5.1.9 / Page 18 / Line 36
Notes that the description for std user categories are their
names and they are not localized.
Resolution: Not Accepted
====================================================================
200 - Clarify response to ignoring security request
Section 5.1.11 / Page 21 / Line 1 / Andre Kramer
Specifies what Producer should do when security requests are
ignored.
Resolution: Not Accepted
====================================================================
201 - PortletDescription applies to clones
Section 5.1.11 & 6.3.3 / Page 21 / Line 29 / Andre Kramer
Proposes which fields from PortletDescription are required to
apply to clones as well.
Resolution: Accepted - Metadata for the portlet applies to
the clone as well.
====================================================================
204 - Fields for cache key
Section 6.1.2 / Page 27 / Line 33 / Andre Kramer
Proposes several additional fields also be part of cache keys
Resolution: Not Accepted - Will deal with these suggestions in
the primer. Alan Kropp will take a look at helping
Gil with the primer.
====================================================================
141 - Add previous windowState and mode?
Section 6.1.2 / Page 26 / Line 46
This would place the burden of remembering the previous
windowsState and mode on the Consumer and thereby permit the
Portlet to know about interesting transitions that do not
involve user interactions with the Portlet's markup.
Resolution: Deferred (pending JSR-168)
====================================================================
205 - Template requirements
Section 6.1.6 / Page 29 / Andre Kramer
Proposes explicit language for how templates may be dependent on
markupType.
Resolution: Not Accepted
====================================================================
206 - Values for extensible items such as modes
Section 6.1.9 and 7.1.1 / Page 31 / Line 13 / Andre Kramer
Consumer vs custom modes and window states
Resolution: Accepted - Rich will look at adding a section
to describe the process for setting the mode.
====================================================================
Discussion on schedule
- A bit behind schedule, but there does not appear to be many
Change Requests where there would be much contention over
the premise.
Mostk of change requests relate to changes in
conformance statements - These fall into 3 major categories.
1) There is one batch of statements that are not clear
when extracted out of context. The CRs related to these
make the conformance statements clear when extracted.
2) Another batch of conformance statements are not testable.
The CRs related to these statements are changed to
semantic language, still capturing the intent of the
original statement.
3) A third batch deals with conformance statements that
are repeated multiple times. For the CRs dealing with
these, the strongest statement was identified, and the
others were dropped down to semantic language. In some
cases the statements got out-of-sync and the differences
needed to be resolved.
It was agreed that all conformance statement related CRs
would be accepted as a group, and Rich would incorporate these
changes in the next version of the spec, using change bars to
indicate the changes.
This will leave 6 outstanding Change Requests.
Rich will create a new version of the spec by next week, with
a target of Tuesday.
====================================================================
Change Requests that remain open:
138 - How does info get to proxied resources?
141 - Add previous windowState and mode?
207 - PortletDescription dependent on UserContext?
213 - Require modes, windowstates & userAuthentication are URIs
215 - Delete statement about 'private conversational state'
230 - Setting navState
====================================================================
Next Meeting:
Thursday March 13
8:00 am PST / 11:00 am EST / 5:00 pm CET
duration: 2 hours
====================================================================
Meeting adjournded at 10:05 AM PST