Get documents about "Meeting Minutes for OASIS SDD_ 2008
Document Sample
Minutes for OASIS SDD TC 09/10/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from September 3.
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Resume review (begun last week) of the collection of pending
specification updates resulting from all prior approved proposals and
discussions (Merri)
B. Discussion Items:
i. Element figures/artwork
ii. id attribute in AuthorizationConstraintType
iii. Request for assistance: specification text for previously
approved proposals.
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern
MINUTES
We have a quorum.
1. Minutes from September 3 approved without objection.
2. Administrative:
a. None new.
3. v1.0 status
a. None new.
4. v1.1 items:
a. Continued review of collection of pending specification changes. Items below
are with respect to the draft specification that Merri has posted in “Working
Documents” on the TC page (v1.1 spec draft, Revision 3).
i. General: Resolved to use the “new new” Oxygen diagrams as the
element illustrations throughout the specification. Merri will make
these updates (thanks, Merri!).
ii. General: Agreed that we need a “deep dive” into the specification to
ensure that we find and update all the places that need to be updated
as a result of any deprecations in v1.1 (we already approved
deprecating the ResourceType name property, and will have a
proposal to deprecate deploymentDescriptor
descriptorLanguageBundle property).
iii. Line 598, Topology: Yes, this statement is still true; leave as is.
iv. Lines 793 – 795, DescriptorInfoGroup: we generally agreed that this
statement reflected our previous discussion; however, Randy prefers
removing descriptorLanguageBundle from deploymentDescriptor
entirely (deprecate this property), and removing it from the schema,
which would remove the need for this clarifying statement. It was
noted that this type is a shared type (in common.xsd), but it can be
moved out of common.xsd to packageDescriptor.xsd. Resolved that
we would take this action, so this addition to the specification is no
longer needed and can be removed. Action: Randy to generate
proposal for this; Brent will present in Randy’s absence at a future
meeting (note: Randy posted this proposal after the meeting).
v. Line 1172, ResourceType example: This isn’t a good example for
this item, either, so when the new example is generated for the
corresponding section, this example should be replaced also.
vi. Lines 1709 – 1712, Constraints: Change to Line 1709 is good. In
Line 1710, added text is good. Line 1710 – 1711, change the last
sentence to simply “Examples are:”, and in Line 1712, add “values” at
the beginning of that bullet.
vii. Line 1719 Constraints: OK as is.
viii. Lines 1958 and onward, AuthorizationConstraintType: Brent to
suggest minor re-wording of first sentence (Lines 1961 – 1962). Re-
word example to be from viewpoint of SDD author rather than
runtime (add “… the SDD author can declare that root authority is
required”).
Question raised about id property: This property does not exist in any
other Constraints, should it be removed? Also, should add description
and shortDescription properties to AuthorizationConstraint to be
consistent with other Constraints. Action: Randy to review notes and
generate proposal about defining or removing id and adding
description/shortDescription.
ix. Line 2056, ConditionalResourceConstraintType: we believe that
AuthorizationConstraintType probably should be added here; will
consider this at the same time we reconsider the preceding items for
AuthorizationConstraintType.
x. Line 2072, ConditionalResourceConstraintType property summary
& usage notes: In addition to the issue noted in the comment, the
description of Name in the summary table does not match the
description of Name in the usage notes. Clearly something needs to
be changed here; will consider this further later. Action: Jason to
generate an example that uses this Type to help determine what is
correct and what needs to be corrected.
xi. Line 2329, ResourcePropertyType: Change “may not be sufficient
enough” to “is not sufficient”.
xii. Line 2438 – 2439, RequirementType property usage notes: Change
“SDD” to “deployment descriptor” (both occurrences); remove
“individual”. Action: Brent to propose text and location for
corresponding content in the composite/aggregated SDD section.
(Update after meeting: Brent doesn’t think another corresponding
note is needed; the Aggregation Requirements section refers readers
back to this section).
xiii. Lines 2485 & 2491: Good as is.
b. We will resume the specification review in TWO weeks.
5. Discussion items:
a. None new
6. Old/new business
a. NO MEETING NEXT WEEK. Will continue reviewing the remaining items in
TWO weeks.
Adjourned 11:50 US eastern time.
Minutes for OASIS SDD TC 09/03/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from August 20 (no meeting August
27).
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Review the collection of pending specification updates resulting
from all prior approved proposals and discussions (Merri)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern
MINUTES
We have a quorum.
7. Minutes from August 20 approved without objection (no meeting August 27).
8. Administrative:
a. Brent reminded everyone of the changes to the OASIS TC process that were
shared with all members via e-mail.
9. v1.0 status
a. None new.
10. v1.1 items:
a. Review of collection of pending specification changes. Items below are with
respect to the draft specification that Merri has posted in “Working
Documents” on the TC page.
i. Line 56, copyright: Brent believes that we should keep the copyright
as originally specified (2007, 2009). Action: Brent will check OASIS
policy and share with TC.
ii. Line 1004, contentType pathname: change “you” to “the SDD
author”.
iii. Line 1023 contentType descriptorLanguageBundle: OK.
iv. Line 1154 ResourceType: Discussion about different usage of Name
property between Topology and Resulting Resource. Confirmed our
previous resolution to deprecate Name property and add text here
that indicates that whatever property, if any, is used as the identifier
in Topology must match that used in ResultingResource. Action:
Jason to suggest the specific text for this.
v. Line 1196 ResourceType name: This property will be removed
(deprecated). Action: Brent to determine any OASIS policy/practice
for documenting deprecated properties. [Added after meeting: have
this information. Suggest leaving this section in, perhaps graying it
out, and adding text at the beginning of this item to indicate that this
property is deprecated in v1.1]
vi. Line 1381, installableUnitType targetResourceRef: suggest a
different example. Action: Jason/Jeff to suggest specific text for a
different example.
vii. Line 1705, Constraints: the added text is OK, but want to change the
next sentences:
1. Line 1706: remove “a property of”
2. Line 1707” change “different types of resource properties” to
“different types of constraints”
3. Line 1708: change “that the property value be” to “such as
values that are”
viii. Line 1842, VersionConstraintType: change “At run-time” to “During
deployment”
ix. Line 1851, VersionConstraintType: added text is OK, but change
line 1851 “resource version” to “version of the resource”
b. Merri also raised the issue that two newer versions of Oxygen produce
different diagrams than the ones in the current v1.0 specification and we
need to decide (1) which version to use for new diagrams; and (2) if we want
to replace all existing diagrams with those created by a newer version of
Oxygen so that all diagrams are consistent. Action: Merri to post examples
of diagrams produced by each version so that TC can decide these questions
[Added after meeting: Merri has posted the examples; please comment by e-
mail and we will discuss next week also]
We will resume the specification review next week beginning on line 1955,
and also discuss the artwork issue noted earlier and other items (Merri has a
couple additional discussion items).
11. Discussion items:
b. None new
12. Old/new business
a. Will continue reviewing the remaining items next week
Adjourned 11:55 US eastern time.
Minutes for OASIS SDD TC 08/20/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from August 13.
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Continue discussion of a collection of specification clarifications
and other related items (Jason & Merri)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern
MINUTES
We have a quorum.
13. Minutes from August 13 approved without objection.
14. Administrative:
a. None new
15. v1.0 status
a. None new.
16. v1.1 items:
a. Continued discussion of collection of minor items, mostly specification
clarifications (Jason & Merri): Have been keeping notes on questions &
feedback based on Eclipse COSMOS implementation experience. Items below
are with respect to the published v1.0 standard at http://docs.oasis-
open.org/sdd/v1.0/os/sdd-spec-v1.0-os.pdf and the draft specification that
Merri has posted in “Working Documents” on the TC page.
i. Line 2312, ResourcePropertyType: Jason: consider a type in
topology that could have multiple instances (e.g., databaseSystem
that could have Oracle and DB2 instances). In this
ResourcePropertyType, I’m looking for database name; the
resourceRef refers to databaseSystem, which doesn’t allow you to
refer to the individual instances of databseSystem. Randy: correct;
can’t resolve resource’s property until the resource is instantiated.
This is a name resolution issue – it’s the value of the property that
we’re interested in. Don’t this we can resolve variables until targeted.
After a good deal of discussion, Resolved that the intended use
doesn’t match the COSMOS expectation. Merri took notes to add
some clarification information to the specification; Jason will update
the COSMOS example and take the results of this discussion back to
the COSMOS team.
ii. Line 2351 Requirements: similar to previous item. Suggest that
information be added to the specification for SDD authors about what
is expected/best practices regardless of the runtime; individual
runtimes can document their behavior/expectations. Merri has some
clarification information to add to the specification.
iii. Line 2716 ResultingResourceType: Resolved that we will strike the
sentence that says “Name SHOULD be defined in Topology…”,
through the end of that paragraph.
iv. Line 3863 TranslationKey: this relates to the
descriptorLanguageBundle issue discussed in a previous call.
Resolved that we don’t need to repeat correct use of
descriptorLanguageBundle here, but we should refer back to that
section where we are adding new guidance.
b. This completes the list of issues we had on the table! Merri will make draft
specification updates for all of the open/pending specification items and
circulate a draft that we will review at next week’s call.
17. Discussion items:
c. None new
18. Old/new business
a. Will continue reviewing the remaining items next week
Adjourned about noon US eastern time.
Minutes for OASIS SDD TC 08/13/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from July 16, 2009 (no quorum to
approve these on July 30) AND July 30, 2009(no meeting July 23, no
formal meeting August 6).
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Continue discussion of a collection of specification clarifications
and other related items (Jason & Merri)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern
MINUTES
We have a quorum.
19. Minutes from July 16 and July 30 approved without objection.
20. Administrative:
a. Brief discussion about TC membership. Currently active organizational
members of SDD TC are CA, IBM, SAP, SAS. Jason indicated that Eclipse is
doing a membership drive and that there might be overlapping interest in the
SDD COSMOS project and the OASIS standard, so we might be able to
leverage the Eclipse membership drive to identify and attract potential new
SDD TC members. Brent to send Jason a list of potential candidate
organizations which Jason will take forward to Eclipse as appropriate.
21. v1.0 status
a. None new.
22. v1.1 items:
a. Continued discussion of collection of minor items, mostly specification
clarifications (Jason & Merri): Have been keeping notes on questions &
feedback based on Eclipse COSMOS implementation experience. Items below
are with respect to the published v1.0 standard at http://docs.oasis-
open.org/sdd/v1.0/os/sdd-spec-v1.0-os.pdf and the v1.1 draft specification
that Merri has posted in “Working Documents” on the TC page.
i. Line 1194, Name property for ResourceType: Question revolves
around use of a resource Name in Topology. Randy’s opinion is that
Name shouldn’t be in Topology; it’s an identifier for a resource and
isn’t needed in multiple places. Jason indicated that this is a
“convenience” property but having it introduces some ambiguity.
After some discussion, Resolved that we will deprecate Name in
ResourceType and add guidance about resource identification to the
effect that a resource needs an identifier but it is not required to be in
a property called Name. Within Topology, we will identify a Property
that identifies the resource. This property might be, but is not
required to be, called Name. If so, it corresponds to the resource
Name property; if not, the Profile should map the resource identifier
to the property (for example, the property that identifies the resource
might be called Label or Tag or Title, in which case the Profile would
specify this equivalence mapping of the property that identifies the
resource). Merri will generate updates to capture this idea as
guidance in the specification.
A good follow-up discussion (not captured here) about usage of such
properties ensued, with a general consensus that Jason will capture in
updates to the SAS SDD example.
23. Discussion items:
d. None new
24. Old/new business
a. Will continue reviewing the remaining items next week
Adjourned noon US eastern time.
Minutes for OASIS SDD TC 07/30/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from July 16, 2009 (no meeting
July 23).
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Continue discussion of a collection of specification clarifications
and other related items (Jason & Merri)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern
MINUTES
We did NOT achieve a quorum.
25. Minutes from July 16 approval deferred until we have a meeting with a quorum.
26. Administrative:
a. Informational: Kirk Wilson indicates that he will be retiring from CA as of
tomorrow.
27. v1.0 status
a. None new.
28. v1.1 items:
a. Continued discussion of collection of minor items, mostly specification
clarifications (Jason & Merri): Have been keeping notes on questions &
feedback based on Eclipse COSMOS implementation experience. Items below
are with respect to the published v1.0 standard at http://docs.oasis-
open.org/sdd/v1.0/os/sdd-spec-v1.0-os.pdf . Today we reviewed the changes
already made in response to the discussion two weeks ago. We deferred
discussion on remaining items until we have a quorum.
i. Lines 779 - 792, descriptorLanguageBundle: change SHOULD to
MUST (this sentence implies that a translation file exists). For last
item in comment, add text indicating that descriptorLanguageBundle
should not generally be in the deploymentDescriptor, but if it is, it
must match what is in the packageDescriptor.
ii. Lines 987 – 1004: change “base” to “root”. Add text around line 1022
that refers readers to the SDD example for localization in the
paragraph beginning on line 987.
iii. Line 1375 vicinity: the cited example is a good addition; will add it.
iv. Lines 1847 – 1848: Remove “proper”. Add text indicating that if
Certified is present, it should be checked first; if Certified is not
present, then Supported is used.
v. Line 2381 (table): Need to consider the case when unique ids are
required when aggregating. Add text around line 2408 indicating that
ids must be unique within an (aggregated) deployment descriptor.
Might want corresponding text in the Aggregation section.
29. Discussion items:
e. None new
30. Old/new business
a. Will continue reviewing the remaining items next week (August 6). Neither
Brent nor Merri will be available next week, so next week’s meeting will be a
general discussion/question and answer session about a few items that need
clarification (Jason wants to discuss these with Randy and anyone else who
can make it to the meeting next week). We will resume the formal TC
business of resolving issues and agreeing on specification changes the
following week (August 13).
Adjourned 11:47 noon US eastern time.
Minutes for OASIS SDD TC 07/16/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from May 7, 2009 (no meetings
between then and today).
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Discuss a collection of specification clarifications and other
related items (Jason & Merri)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern time ** Adjourn 12:00 Noon US Eastern
time
MINUTES
We have a quorum.
31. Minutes from May 7 approved without objection.
32. Administrative:
a. Informational: Chris Robsahm indicates that he will be leaving SAP in August.
Chris will investigate SDD interest at his new employer once he is engaged
there. Jason also indicated that he is following up with both SAP and Acresso
contacts to see if there is renewed interest in active participation there.
33. v1.0 status
a. None new.
34. v1.1 items:
a. Collection of minor items, mostly specification clarifications (Jason & Merri):
Have been keeping notes on questions & feedback based on Eclipse COSMOS
implementation experience. Items below are with respect to the published
v1.0 standard at http://docs.oasis-open.org/sdd/v1.0/os/sdd-spec-v1.0-
os.pdf
i. Line 781, descriptorLanguageBundle: suggest adding that there
SHOULD be at least one languageBundle (it’s effectively required if the
intent is to display any property values in a language other than what
the SDD is composed in – this implies that translationKey effectively,
but not strictly, becomes required if any property values are to be
displayed) and deprecate use of inline string and add annotations for
human-readable content (replaces what was intended to go in the
inline strings); location of base package in the resourceBundle goes in
the packageDescriptor. Randy: many other properties (name,
description, several others) also have translationKey. This same
rationale and associated changes apply to all of these properties.
ii. Line 1022, descriptorLanguageBundle: suggest adding text to
indicate the URI of a bundle is also included in the packageDescriptor.
Randy pointed out that these are just enum values. So instead we
probably need to clarify what SHOULD be in the packageDescriptor.
Need a more thorough description of this overall element higher up in
this section to better describe intended usage (suggest additions
around line 721 and/or line 981 – a quick explanation in the
specification with a reference to a new example that shows localized
content in the Primer). Need to think more about the new example
(should it build from an existing example, adding multiple languages,
or be a new example? Jason has a COSMOS example and will send
that to the TC for review).
iii. Line 1847, Certified: unclear what runtime should do with this
information vs. Supported. Randy: Certified typically is a narrower
range than Supported. Default should be to check the version with
respect to the range specified in Supported; optionally, could have a
way to instruct the runtime (via a runtime property), if it supports the
capability, to check with respect to Certified rather than Supported. If
Certified is not present, assume that Supported applies. Will update
specification to reflect this.
35. Discussion items:
f. None new
36. Old/new business
a. Will continue reviewing the remaining items next week. Brent will not be
available but will make sure that the teleconference bridge is available and
we just need someone to take notes during next week’s meeting.
Adjourned 12:00 noon US eastern time.
Minutes for OASIS SDD TC 05/07/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from April 16, 2009 (no meetings
April 23, April 30).
2. Administrative
- None new
3. v1.0 status
- No updates
4. v1.1 updates
A. Localized content proposal (Jason)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
37. Minutes from April 16 approved without objection.
38. Administrative:
a. None new.
39. v1.0 status
a. None new.
40. v1.1 items:
a. Jason posted information – not yet a formal proposal – regarding issues with
localized content and display text. See the slides posted for details. After
some discussion, we achieved consensus among those present on the call
that we can document the desired operation in the specification; no schema
changes are required. Published examples also should be updated to match
the guidance to be incorporated in the specification. Because this is a
specification update only and is not deemed controversial, and because it
seems likely, based on input from several TC members, that we will not
achieve a quorum for next week’s call, we agreed to post an electronic ballot
so that this item can be moved forward without delay. Jason will post an
updated proposal and Brent will open a ballot referring to that proposal.
41. Discussion items:
g. None new
42. Old/new business
a. None new.
Adjourned 11:37 US eastern time.
Minutes for OASIS SDD TC 04/16/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from April 2, 2009 (no meeting
April 9).
2. Administrative
- None new
3. v1.0 status
- Specification erratum that was approved has been published at
http://docs.oasis-open.org/sdd/v1.0/ .
4. v1.1 updates
A. Pending specification updates review (see http://tools.oasis-
open.org/issues/browse/SDD ).
B. Resource resolution model (Jason)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
43. Minutes from April 2 approved without objection.
44. Administrative:
a. None new.
45. v1.0 status
a. The specification errata that we approved last has been published at
docs.oasis-open.org. This item is now complete; Brent has updated Jira.
46. v1.1 items:
a. None new
47. Discussion items:
h. Pending specification updates: Merri suggests that we incorporate the
approved items into a new version of the draft specification, as some of the
items are getting “stale” and should be recorded in the specification. Agreed
that we will do this. Brent will update the list of pending specification items
and Merri will begin to update the specification. We will review the v1.1 draft
at one or more later meetings.
i. Resource resolution discussion: Jason described the current COSMOS
implementation, in which they need the context of the SDD to perform
resource resolution property. They have been re-factoring the resource
resolution model to use the SDD schema directly (for example, parameter
values such as installLocation – user might input the value and the runtime
checks for existence of the directory, that the directory is writeable, etc.).
COSMOS implementation is using the <any> elements of the SDD schema to
store such values. Jason is asking if the SDD schema should be extended to
directly support such function. Howard asks why the values are stored in the
XML tree; could they be cached elsewhere? Jason notes that for more
complex cases, storing the value can prevent the need to re-check the value
repeatedly, but in addition, values might be inter-dependent and using the
SDD schema model offers a structured way to keep the values, and the values
could change over time during the deployment as deployment operations are
executed. Hence, they have chosen to use the SDD schema as the model for
the “working store”. Discussion led to observation that this might be unique
to the COSMOS implementation, not generally applicable, and Howard
suggests not extending the schema and continuing to use the <any>
elements as Jason described. Randy noted that he would prefer not to
correlate the schema with the in-memory or persistent data model for the
runtime. Discussion led to consensus: no need to extend the SDD schema
to accommodate the resolution and storage of these values.
Implementations can choose to use the schema model, including the
<any> extension points, in the way COSMOS is doing this, but there’s no
standardization of this concept that is required.
48. Old/new business
a. Jason’s pick from two weeks ago, for UNC to win the Mens NCAA basketball
tournament, turned out to be correct.
Adjourned 11:37 US eastern time.
Minutes for OASIS SDD TC 04/02/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from March 12, 2009 (no meetings
March 19, March 26).
2. Administrative
- None new
3. v1.0 status
- Specification erratum that was approved has been published at
http://docs.oasis-open.org/sdd/v1.0/ .
4. v1.1 updates
A. Expected this week: "ImplementedBy" proposal (Jason)
** See Jira tool for current status of all v1.1 items
B. Discussion
- As required
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
49. Minutes from March 12 approved without objection.
50. Administrative:
a. None new.
51. v1.0 status
a. The specification errata that we approved last has been published at
docs.oasis-open.org. This item is now complete; Brent has updated Jira.
52. v1.1 items:
a. None new
53. Discussion items:
j. Jason reviewed his “ImplementedBy” proposal (see the slides posted to the TC
Web page). Q: Does this help to achieve a sort of “late binding”? A: Yes, that’s
one use; also useful for pre-requisite checking. Randy: helps deal with
polymorphism when two resources actually represent the same instance
because of constraints in the resource model; enables type-casting and
mapping to a particular physical instance. Q: Any reason that the
ImplementedBy attribute refers to another HostedResource via idref rather
than to a resource type from the profile via qName? A: Originally pursued the
qName approach. Randy: difference is that this actually maps to an instance,
not just to a resource type (could have multiple resources of same type).
After additional clarification discussion, it was agreed that the qName
approach will work and is preferable. Consensus to change ImplementedBy
attribute from idref to qName. Motion Randy, second Mark to approve the
proposal with the change noted (idref qName) approved without
objection (including Chair vote if required). Jason will update the
example to show “multiple resource” case and generate an outline of the
specification text; Randy will update and post the schema; Brent will
update the Jira issue.
54. Old/new business
a. Jason picks UNC to win the Mens NCAA basketball tournament; Randy picks
MSU.
Adjourned 11:37 US eastern time.
Minutes for OASIS SDD TC 03/12/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from March 12, 2009.
2. Administrative
- None new
3. v1.0 status
- Specification erratum that was approved has been published at
http://docs.oasis-open.org/sdd/v1.0/ .
4. v1.1 updates
A. Expected this week:
- None known
** See Jira tool for current status of all v1.1 items
B. Discussion
Intended use of package descriptors as they relate to listing content
for the runtime. For example, the spec says that the content should be
content that is deployed but it's not clear whether we should model the
runtime components in the PD or implement an internal media descriptor
like meta file? (Jason)
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
55. Minutes from March 5 approved without objection.
56. Administrative:
a. None new.
57. v1.0 status
a. The specification errata that we approved last has been published at
docs.oasis-open.org. This item is now complete; Brent has updated Jira.
58. v1.1 items:
a. None new
59. Discussion items:
k. Jason raised a question about intended use of package descriptors (PDs) as
they relate to listing content for the runtime. That is, should runtime
components that are “loaded” during deployment (such as Handlers in the
COSMOS implementation) be described in a PD in a manner similar to the
software that is being deployed? Randy suggests that such runtime
components could be described in their own PD (they have their own identity,
including manufacturer, etc., that may be different from the software that is
being deployed); that PD is then aggregated with other PDs that describe the
actual software that is being deployed. The runtime components can be
referenced as pre-requisites for deploying the software. Preference is to
describe the runtime components in their own PD rather than intermingle
with the PD(s) of the software that is being deployed. Howard suggests that
“primary” or “default” runtime components could be provided via some sort
of convention (registry, well-known location, etc.), with “special” (perhaps
third-party, etc.) runtime components being described in a PD. Jason agrees
that this is the type of situation that led to this discussion. Randy notes that
the suggested packaging generalizes such that runtime components for
multiple runtimes (perhaps COSMOS and several others) could all be supplied
in their own PDs. Brent asks if this drives us to CL2; answer is that it does
not; we can aggregate PDs in CL1. Consensus is that this approach of
providing runtime components in their own PDs seems workable and is
a reasonable way to move forward. Jason will share any implementation
experience gained using this approach and suggest proposals, if any, to
clarify it via examples, specification updates, etc.
60. Old/new business
a. None
Adjourned 11:25 US eastern time.
Minutes for OASIS SDD TC 03/05/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from February 26, 2009.
2. Administrative
- None new
3. v1.0 status
- Specification erratum that was approved last week has been sent to
Mary McRae, OASIS TC Administrator, for publication.
4. v1.1 updates
A. Expected this week:
- None known
** See Jira tool for current status of all v1.1 items
B. Discussion
Demonstration of Eclipse COSMOS SDD software implementation (Jason)
** Join the Web conference:
Topic: COSMOS Demos for OASIS SDD TC
Date: Thursday, March 5, 2009
Time: 11:00 am, Eastern Standard Time (GMT -05:00, New York)
Meeting Number: 711 861 509
Meeting Password: sddrocks
Please click the link below to see more information, or to join the
meeting.
-------------------------------------------------------
To join the online meeting
-------------------------------------------------------
1. Go to
https://sasmeetings.webex.com/sasmeetings/j.php?ED=113786362&UID=0&PW=e
3bf5def314749171609095903
2. Enter your name and email address.
3. Enter the meeting password: sddrocks
4. Click "Join Now".
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
61. Minutes from February 26 approved without objection.
62. Administrative:
a. None new.
63. v1.0 status
a. The specification errata that we approved last week has been sent to OASIS
TC Administrator for publication at docs.oasis-open.org.
64. v1.1 items:
a. None new
65. Discussion items:
Demonstrations of the Eclipse COSMOS SDD Build-Time Generator (BTG) and Runtime
code were performed by SAS. In addition to TC members/observers present
(recorded on TC Web page), invited guests from member companies attended to
view the demonstrations: Acresso: Jiju Jacob; CA: Brad Beck; IBM: Stan Spurgeon, pat
Rago, Chris Mildebrandt, Perry Statham, Joe Loewengruber, Ron Doyle, Erich Magee;
SAS: Michael King, Josh Hester. Jason provided a presentation about the demo,
followed by the BTG and Runtime demonstrations. The presentation will be posted
on the TC Web page. Rather than attempt to describe the demo that was shown,
these minutes reflect the questions and answers during the demonstrations.
l. Build-Time Generator:
Q: Is the Management Data Repository (MDR) from the CMDBf standard? A:
Yes
Q: Is this limited to CL1? A: Current COSMOS implementation is; the demo
today shows SAS’s implementation based on COSMOS.
Q: Where is the input for BTG? It comes from a Reader; the demo uses
published OASIS SDD examples. Could also read from a database or could
develop readers for other sources (e.g., RPM).
Q: What is the use case for the SDD merge demonstrated? A: Have two
separate SDDs for individual products/components that we want to
consolidated into a single, composite “solution” SDD.
Q: Is there a use case for people not familiar with SDD, just to “wrap” existing
build information/artifacts in SDD metadata? A: That’s exactly what the
Readers do; they’re the customizable pieces; you can build readers for your
“raw” artifacts (for example, have one for RPM artifacts now).
m. Runtime:
Q: Is P2 the update management system of Eclipse? A: Yes – P3 provides
installer application for Eclipse; SAS adapted it for this demo purpose.
Q: Is installer separate from runtime? A: Yes, installer provides the UI here;
the runtime gnerates the deployment model and lays down the software.
Q: So the installer GUI can be easily replaced with something else? A: Yes,
that’s the design intent.
Q: Are the Resolvers implemented in all Java? A: Some Java, some native (JNI).
So far have a core set of operating system, file system and directory
Resolvers.
Q: Is all this pluggable? A: Yes, one provides a pluggable Handler for the
Profile.
Q: We had an issue in the TC, dealing with the standard, about showing
progress indication. The method shown in the demo works well because of
tight coupling between the handler events and the installer/UI, but that
might not always be the case. A: It is up to the Handler to define and publish
events and the installer to catch and interpret events, so it does require some
common understanding. Follow-up Comment: But when deploying with
legacy artifacts, it’s unlikely that we’ll have this ability for such granular
progress indication. Response: Agree.
Q: Is there a well-defined Handler interface? A: Yes.
Q: What about the deployment “interview” process on installation (e.g., for
verifying disk space, installation directory, port numbers, etc.) – especially for
substituting values for variables in the SDD? Can we interrogate the runtime
during the interview process/can I write my UI and leverage the runtime
capabilities, rather than duplicating all the checks/validations? A: Today,
runtime makes a first pass through the SDD to find “default” resolutions.
Ultimately, SDD variables need to be surfaced to the UI; need a mechanism to
do this; not there yet but requirement is recognized.
Q: Have we thought about this bidirectional interface? A: Information is in the
resolution model, but haven’t leveraged all of it and haven’t completed
interface definition/design, but it’s on the to-do list. Comment: Need to
study this from both implementation and Profile perspectives, especially for
interoperability considerations.
66. Old/new business
a. None
Adjourned 12:02 US eastern time.
Minutes for OASIS SDD TC 02/26/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from February 19, 2009.
2. Administrative
- Everyone's Jira account should be active. We will use Jira for issues
tracking.
- Notice: For next week, we plan to have demonstrations of COSMOS
projects
3. v1.0 status
- Specification erratum 15-day public review began 02/03/2009 and
concluded 02/18/2009.
*** Need vote on final approval of this Errata so that it can be
published (must pass by Full Majority vote)
4. v1.1 updates
ACTIVE PROPOSALS (review/discuss/take action on available proposals as
time permits). See spreadsheet "v1.1 Candidate Items" (and/or Jira
tool) for details
A. Expected this week:
- None known
** In process, tabled, completed proposals no longer included in agenda
(See spreadsheet "v1.1 Candidate Items" and/or Jira tool for details)
B. Discussion
- Profile updates, if any (Jason)
- Examples updates (Randy) *** Need vote to publish updates
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
67. Minutes from February 19 approved without objection.
68. Administrative:
a. Noted that next week we plan to have demonstrations of some of the Eclipse
COSMOS SDD implementation that is under development. Brent to ensure
that Chris knows about this, as he had requested this at a previous call.
Jason to set up WebEx Web conference for the demos.
69. v1.0 status
a. OASIS 15-day public review process for specification errata began February 3
and concluded February 18 with no comments. According to the OASIS
process:
(d) After the public review, confirming the proposed corrections as Approved
Errata by a Full Majority Vote.
We confirmed that there were no public review comments for the errata.
Merri Jensen made a motion for final approval of the specification errata;
Randy George seconded; Mark, Jason, Merri, Jeff, Randy and Howard all
voted Yes. 6 yes votes constitutes the Full Majority required to pass
(with Brent also indicating that his Chair vote, if required, is Yes); motion
approved.
70. v1.1 items:
a. None new
71. Discussion items:
a. Examples update: Randy posted updated examples with changes to the
Monolithic and Language Pack examples that mirror the changes made last
week to the JRE example (generally, moving elements from Topology to
Requirements). Jason also made corresponding updates to the COSMOS
example that he had previously distributed with the Profile. Jason suggested
including that COSMOS example in our published examples. Updated
examples approved without objection, including the new COSMOS
example, for OASIS publication as a new version (expository documents).
It was also noted that these updates will require a small update in the
specification. Randy/Jason will identify the area of the specification to
be updated and we will address this in a future meeting.
b. Profile updates: Jason recapped last week’s discussion about using software
name/version as “normalized” software identifiers and the best way to
specify this identity “mapping” for ResultingResource information. Jason
proposes that this need not be done in the Profile and is best done in the
explanatory document that accompanies the profile and/or the specification.
The distilled version of the practice suggested is: use “generic” resource in
cases that don’t map directly to the resource model (e.g., CIM); use the
Resource Model (e.g. CIM) resource types for cases where the mapping is
direct/straightforward. TC consensus is to proceed according to Jason’s
recommendation. Jason will update the explanatory document that
accompanies the profile for a future meeting.
72. Old/new business
a. None
Adjourned 11:31 US eastern time.
Minutes for OASIS SDD TC 02/19/2009 Meeting
Agenda:
. Minutes approval: Meeting Minutes from February 5, 2009 (no meeting
February 12).
2. Administrative
- We now have a JIRA account and OASIS TC Admin has set up user
accounts for all TC members; you should have received e-mail with
account information. Please test this out and let Brent know your
results.
3. v1.0 status
- Specification erratum 15-day public review began 02/03/2009 and
concluded 02/18/2009.
- Updated examples approved (Brent sent request to OASIS Staff to re-
publish)
4. v1.1 updates
ACTIVE PROPOSALS (review/discuss/take action on available proposals as
time permits). See spreadsheet "v1.1 Candidate Items" (and/or Jira
tool) for details
A. Expected this week:
- None known
** In process, tabled, completed proposals no longer included in agenda
(See spreadsheet "v1.1 Candidate Items" and/or Jira tool for details)
B. Discussion
- Profile updates (Jason)
- JRE Example update (Randy)
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
73. Minutes from February 5 approved without objection.
74. Administrative:
a. None new
75. v1.0 status
a. OASIS 15-day public review process for specification errata began February 3
and concluded February 18 with no comments.
76. v1.1 items:
a. None new
77. Discussion items:
a. Profile updates: Jason posted updated documents to the TC Web page. Main
items essentially the same as previous revision; updates this time include
adding all types from v1.0 Profile that cover our v1.0 published examples
and new Artifact and AdditionalContent items per previous discussions.
Motion by Jason, second by Randy to approve this version as the new
Profile Schema. Discussion: Clarified that the Profile is a non-normative
(expository) document. Question from Jason about IPR for such OASIS
publications with respect to Eclipse use of SDD Profile schema; Brent
explained his understanding that OASIS deliverables are intended to be used
openly and freely, including expository documents; that our TC operates
under RF on Limited Terms OASIS IPR policy, with no known IPR statements
or claims from any party, so use by Eclipse or other parties should be
permissible. Brent to follow up with OASIS staff to confirm or correct this
understanding. Motion approved without objection; Jason will develop
and submit an explanatory document for the Profile (similar to the one
for v1.0) to accompany the approved Profile schema.
b. Examples update: Randy posted updated examples with changes to the JRE
example that are driven by the new Profile. Several items were “moved” from
Topology section to Requirements section; Randy believes this makes for a
cleaner distinction and illustrates a best pratice of using Topology for “type-
casting” only, and might present a future direction to remove all Variables
from Topology (not proposing this now, just an observation). Consensus is
that this is the right approach; Randy will determine if other examples
need similar updates (Monolithic example does); Jason will make
corresponding updates to the example he included with the Profile;
Randy will propose specification change to address this best practice
and might propose a schema change to enforce it.
78. Old/new business
a. Jason asks about use of ResultingResource information, such as
name/version in the ResultingResource and how these relate to name/version
of the “same” software elsewhere in the SDD. Randy notes that name/version
is used as a normalized software “identifier” of sorts; and the information in
ResultingResource is not expected to differ from what is specified for the
“same” software elsewhere in the SDD. Randy notes that this “mapping” of
identifier information to SDD properties should be specified in the Profile.
Jason agrees and will propose an addendum to the new Profile to accomplish
this.
Adjourned 11:48 US eastern time.
Minutes for OASIS SDD TC 02/05/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from January 15, 2009 (no meetings
January 22 or January 29).
2. Administrative
- We now have a JIRA account and OASIS TC Admin has set up user
accounts for all TC members; you should have received e-mail with
account information. Please test this out and let Brent know your
results.
- Determine whether or not to meet next week (Brent will not be
available)
3. v1.0 status
- Specification erratum to update pointers to schema files: TC approved
Oct. 30; Brent sent errata package to OASIS Staff 12/01/2008; 15-day
public review began 02/03/2009 and concludes 02/18/2009.
- Updated examples approved (Brent sent request to OASIS Staff to re-
publish)
4. v1.1 updates
ACTIVE PROPOSALS (review/discuss/take action on available proposals as
time permits). See spreadsheet "v1.1 Candidate Items" (and/or Jira
tool) for details
A. Expected this week:
- None known
** In process, tabled, completed proposals no longer included in agenda
(See spreadsheet "v1.1 Candidate Items" and/or Jira tool for details)
B. Discussion
- Profile updates (Jason)
- COSMOS update (Jason)
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
79. Minutes from January 15 approved without objection.
80. Administrative:
a. Jira accounts set up – several on the TC have successfully logged in. Request
that remaining members test their accounts – if you do not have your
account information, contact Brent. We will migrate to using the Jira tool for
our item/issue tracking starting now.
b. Plan to cancel next week’s meeting (both Randy and Brent will be out), unless
items are raised. Jason expects to have a Profile proposal the following week.
81. v1.0 status remains as noted in agenda
a. Brent has asked OASIS staff to publish our updated examples (with correct
URLs) that were approved previously.
b. We approved the specification erratum to update pointers to the schema files
in the specification; the OASIS 15-day public review process began February 3
and will conclude February 18.
82. v1.1 items:
a. None new
83. Discussion items:
a. Profile updates: Jason posted updated documents to the TC Web page. Added
a “hosts” relationship for each resource. Randy: past experience indicates
that some software could have multiple roles – e.g., a database server or an
application server can themselves be applications, but we might need to
know their specific type of application. Need to explore this relationship
further to ensure we cover all the bases in this area. PropertyProvider
element has been added with “affinity” attribute (addresses previously
identified issue about “intertwined” properties – e.g., OSType and OSVersion
– by specifying that there is some connection between/among properties)
and “handler” attribute (URI reference to class/jar/executable/etc. that
handles this particular resource). Randy: isn’t affinity inherent? Jason: sort of,
with semantic understanding of the properties, but nothing available that
informs the runtime about ordering of property processing. Affinity is
optional and enables “good SDD authors” to specify affinity that implies
ordering of processing. This direction is that the SDD schema will not specify
this information; it will be in the Profile; after some discussion, it seems we
have general consensus that this (Profile, not schema) is the preferred
method. Remaining open items for the profile include the resource handler
coupling and what can be defined in AdditionalContent (what types the
runtime can handle). These will be added to the proposal that will be brought
forward in the future (target next meeting, in 2 weeks). Randy: Do we need to
add any additional guidance/stronger statements to the specification
regarding use of Name property? Jason: Specification is pretty clear now; the
JRE example kind of takes us in the wrong direction. Resolved that we will
examine and update this example and any others that might have the
same issue (Randy) and we’ll review changes at next meeting in two weeks.
b. COSMOS update: Build-time generator for CL1 complete and available.
Runtime code is continuing development – converted from Tuscany to JAXB
implementation; a change resolver is now operational. v1.1 iteration 2 is now
in test; iteration 3 is the milestone driver and is expected to be available
April 3; goal is to have basic runtime framework in place that can install
COSMOS itself. Overall v1.1 release target is June.
84. Old/new business
a. Jason notes that there are 3 COSMOS projects (Build-Time Generator,
basically done); Runtime (underway); and SDD editor (not getting much
attention now; Acresso or others might be interested in this). Jason
encourages any interested parties to become involved, especially in the SDD
editor project.
Adjourned 11:56 US eastern time.
Minutes for OASIS SDD TC 01/15/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from January 8, 2009
2. Administrative
- We now have a JIRA account. Brent has move operational items onto
this new system, but OASIS TC Admin needs to set up user accounts for
all TC members before we can use this system.
3. v1.0 status
- Specification erratum to update pointers to schema files: TC approved
Oct. 30; Brent sent errata package to OASIS Staff 12/01/2008 for the
required 15-day public review.
- Updated examples approved (Brent sent request to OASIS Staff to re-
publish)
4. v1.1 updates
ACTIVE PROPOSALS (review/discuss/take action on available proposals as
time permits). See spreadsheet "v1.1 Candidate Items" (and/or Jira
tool) for details
A. Expected this week:
- None known
** In process, tabled, completed proposals no longer included in agenda
(See spreadsheet "v1.1 Candidate Items" and/or Jira tool for details)
B. Discussion
- General discussion of going-forward plan (v1.1 with CL1 updates,
migration to new Profiles, v1.2 with CL2 updates, any F2F meetings,
etc.)
5. Old/new business
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
85. Minutes from January 8 approved without objection.
86. Administrative:
a. Mary McRae from OASIS just set up a JIRA account for us. Brent has entered
the items/issues currently being tracked in the spreadsheet (including
completed items). Mary is setting up accounts for all TC members; you
should receive e-mail notification from OASIS once everything is set up.
87. v1.0 status remains as noted in agenda
a. Brent has asked OASIS staff to publish our updated examples (with correct
URLs) that were approved previously.
b. We approved the specification erratum to update pointers to the schema files
in the specification; Brent has transcribed this information into the OASIS
document template so that this can go through the OASIS 15-day public
review process. Mary McRae from OASIS is initiating this public review.
88. v1.1 items:
a. Confirmed last week’s direction: We resolved that we will focus on CL1 for
the v1.1 specification, with proposals primarily generated from
implementation (notably COSMOS) experience (although anyone may
make any proposal at any time, as always), with target to issue v1.1
specification in the summer; and then produce another version (probably
v1.2) with any CL2 updates (many of the current items in the
spreadsheet are CL2) – v1.2 not scheduled at this time.
b. We also agreed that we will issue the new Profile(s) once the COSMOS profile
implementation completes (likely to be a little bit before a v1.1 specification).
c. Merri Jensen agreed to act as specification editor for v1.1; Brent will assist.
Most specification updates are relatively minor and have already been
developed within the approved proposals. Thanks Merri!
d. Face-to-face meeting: none planned at this time; if v1.1 updates merit an in-
person discussion, we will consider a meeting at that time.
89. Old/new business
a. Jason brought up an issue discovered during COSMOS implementation:
sometimes an SDD specifies multiple properties per resource, where some
properties are “dependent” on others. For example, OSType and Version
might both be specified, but OSType must be resolved first, because the
OSType determines how to obtain the version (a process that varies across
operating systems). Schema doesn’t indicate these dependencies; question
is, should we do anything in the specification/schema to address this item?
Howard: wouldn’t an implementation plug-in solve this? Jason: Yes, that’s
one way to solve, by leaving this item to the implementations. Randy:
Topology doesn’t provide the necessary information. Could use dependency,
making it richer, for CL2 cases. In CL1, multiple properties on a resource are
logically ANDed together; implementations plug-in should handle any
ordering within a single resource. Can’t think of a particular example with
dependent properties across resources. After discussion, consensus is that
implementation handles ordering of property resolution; no change in
specification schema. As we go forward, can consider author declaration of
known dependencies and consider specification/schema/Profile/Primer/Best
Practices updates if justifying use cases arise.
Adjourned 11:53 US eastern time.
Minutes for OASIS SDD TC 01/08/2009 Meeting
Agenda:
1. Minutes approval: Meeting Minutes from November 20, 2008
2. Administrative
- We now have a JIRA account. Brent will proceed to move operational
items onto this new system
3. v1.0 status
- Specification erratum to update pointers to schema files: TC approved
Oct. 30; Brent sent errata package to OASIS Staff 12/01/2008 for the
required 15-day public review.
- Updated examples approved (Brent sent request to OASIS Staff to re-
publish)
4. v1.1 updates
ACTIVE PROPOSALS (review/discuss/take action on available proposals as
time permits). See spreadsheet "v1.1 Candidate Items" for details
** Expected this week:
- None known (will review list of pending items during the meeting)
** In process or tabled (to resume discussion later):
- D2 Self-Documenting Schema (Howard): proposal direction was approved;
seeking volunteers to develop final proposal
- S3 Interoperability (Julia): Tabled until Julia returns; may want to
split this into separate items addressing various aspects of
interoperability
COMPLETED PROPOSALS (for reference):
- S0 Charter update (Brent): Complete/published
- N18 Use OASIS Standard for name/address (Howard): Initial
outline/sketch to determine direction was presented; TC consensus is
not to pursue further.
- N14 Updated/Final Pattern proposal (Randy): approved;
specification/schema updates pending
- N1a Root Install (Randy & Jason): approved; specification/schema
updates pending
- N12: Additional Variable Types (Lazar/Chris): approved;
specification/schema updates pending
- D8: Specification erratum for latest schema files (Brent) Approved;
need to generate documentation in OASIS template format.
- N9/N10: Version representation/Revisit use of Profiles (Jason):
Proposal as amended after discussion was approved; this gives us a new
"foundation" for profiles.
C. Discussion
5. Old/new business
- Brent has action item to schedule discussion with OVF Working Group
(original target end October; new target 2Q 2009)
- Brent has action item to schedule discussion with OASIS SCA-Assembly
TC (request sent to SCA-Assembly co-chairs 08/21; no reply; sent
reminder 09/24; no reply)
** Adjourn 12:00 Noon US Eastern time
MINUTES
We have a quorum.
90. Minutes from November 20 approved without objection.
91. Administrative:
a. Mary McRae from OASIS just set up a JIRA account for us. Brent will enter the
items/issues currently being tracked in the spreadsheet, and the pending
specification items, into Jira and we’ll use this tool until further notice. Brent
will send out access information to the TC once everything is set up.
92. v1.0 status remains as noted in agenda
a. Brent has asked OASIS staff to publish our updated examples (with correct
URLs) that were approved previously.
b. We approved the specification erratum to update pointers to the schema files
in the specification; Brent has transcribed this information into the OASIS
document template so that this can go through the OASIS 15-day public
review and approval process. Mary McRae from OASIS advises that this public
review should commence this week.
93. v1.1 items:
a. No new items this week; we discussed the existing high-priority items in the
spreadsheet, and Jason noted that as the COSMOS iterations proceed, items
that might require proposals for the SDD standard are likely to arise
(although the specific line items aren’t yet generated). After examining the
high-priority items in the spreadsheet, it became apparent that we generally
will be better off thinking about specification improvements in the context of
implementation experience. We did not identify any items that anyone
believes need immediate proposals, or that anyone wants to immediately
generate proposals for. We also noted that the COSMOS implementation
focuses on CL1, and the items we have already approved for v1.1 also are
restricted to CL1. Hence, we resolved that we will focus on CL1 for the
v1.1 specification, with proposals primarily generated from
implementation (notably COSMOS) experience (although anyone may
make any proposal at any time, as always); and then produce another
version (probably v1.2) with any CL2 updates (many of the current items
in the spreadsheet are CL2).
94. Old/new business
a. Joint discussion with SCA-Assembly TC (whose charter notes that they will
consider leveraging SDD): Brent has contacted the SCA-Assembly TC co-chairs
requesting a meeting with that TC to review/provide education on SDD. No
response to that request yet. Brent followed up with another query on 09/24;
no response to that yet, either.
b. Joint discussion with OVF working group: Brent will make contact with OVF
representatives in 1Q 2009, based on information received from that group
as to when they reach an appropriate milestone that would present an
opportunity for such a discussion..
Adjourned 11:20 US eastern time.
