Minutes for OASIS SDD TC 03 / 22 /2007 Meeting
AGENDA
11:05 AM ET Roll Call
1. Minutes approval Meeting Minutes from February 22, March 8 and March 15 meetings (note: March 1 meeting cancelled; did not have quorum at March 8/15 meetings).
2. Administrative (Julia). Also:
- Ballot open for selecting date for next F2F meeting. Ballot open until March 27; please vote for all of the selections for which you could make it.
3. Custom entities proposal (Julia/Rich) -- this was presented previously but could not be formally voted on owing to lack of quorum.
5. Version proposal (Brent)(if time allows)
6. Old/new business
- continue profile discussion?
- any other items
** 12:00 noon ET -- adjourn
MINUTES
We have a quorum.
1. Minutes for February 22, March 8 and March 15 are approved without objection.
2. Ballot still open on F2F dates. SAP mentions they cannot make it the week of the 23rd. Brent notes that the first week in May is the current leader.
3. Custom “stuff” proposal – continue discussion. Julia recaps proposal based on posted proposal presentation. Randy argues in favor of alternative 1 – removing the current custom check element and documenting these concepts in the spec rather than adding custom logic for other elements that correspond to the environment. We vote and agree to accept alternative 1 presented at the F2F rather than the current “custom stuff” proposal (alternative 2).
4. Version proposal – Brent has posted this in the Working Documents title. Debra notes that the ability to compare versions is conceptually the same as the ability to discover the values of resource properties – a function we have decided to leave entirely up to the implementation. This is an argument against stepping into the very complex arena of version comparison. We discuss and agree to just make version a string and acknowledge that implementations need to understand how to compare versions
Minutes for OASIS SDD TC 03 / 15 /2007 Meeting
MINUTES
We do NOT have a quorum. The unusual daylight savings time change resulted in the SAP folks in Germany calling in late. Chris, Frank H., Lazar and Thomas K. of SAP are credited with attendance.
Minutes for OASIS SDD TC 03 / 08 /2007 Meeting
AGENDA
1. Minutes approval Meeting Minutes from February 22 meeting (note: March 1 meeting cancelled).
2. Administrative (Julia). Also:
- Ballot has opened for selecting date for next F2F meeting. Ballot open for 3 weeks; please vote for all of the selections for which you could make it.
3. OASIS Symposium update from Jane Harnad, OASIS (5 min.)
4. Custom entities proposal (Julia)
5. Version proposal (Brent) (if time allows)
6. Old/new business
- continue profile discussion?
- any other items
MINUTES
We do NOT have a quorum.
1. Minutes from 2/22 cannot be approved without a quorum.
2. Ballot for next F2F is now open. Currently pursuing either Austin (to be co-hosted by IBM and Dell) or Raleigh (co-hosted by IBM and SAS) as the location.
3. Jane Harnad of OASIS: OASIS symposium to be held April 15-20, 2007, in San Diego, Calif. Event details available from URL on OASIS home page.
4. Custom Stuff proposal (Rich and Julia): Rich has posted a presentation in the working documents section. Some comments from a walk through the presentation:
a. All the concepts apply to current artifacts; nothing new being introduced.
b. Q: How to separate (for example) lifecycle operation artifacts from resulting resources?
A: The timing, inputs, and functions are different.
c. Q: Does the presentation list all the areas affected by the proposal?
A: The list on slide 3 ("SDD elements that correspond to the deployment environment") was generated from Julia's going through the entire schema; the table on slide 5 is exhaustive. The left column is the scope; the rows are grouped by function; input is the base type name.
d. "CustomStuff" is obviously not the final name. Brent suggests "CustomItems" as a stopgap to keep us from getting too comfortable with "stuff".
e. Slide 8 (CustomStuff element in schema): there's an implied maxOccurs="1" (which is explicitly displayed in the illustration on the previous slide).
f. Slide 9 ("Supporting Types"): AllElementsSupported should be removed. Also, only appropriate element types supported in each element of type CustomArtifactType; not enforced in schema, but to be described in spec.
g. Much discussion of the semantics of having this in SDD: does it mean "must be run" or "should be run"? Suggestion that if it's optional, should be called "hints".
h. Is custom stuff an override? Only makes a difference if there are side effects; discovery, for example, should not have side effects.
i. There is concern that the same semantics cannot be used for all the types of custom stuff being discussed. To revisit.
j. Slide 13 ("Implementation Implications"): there is concern about interoperability. Does the implementation decide whether a custom artifact is performed once or multiple times if associated with multiple elements?
k. Discussion about the value of the SDD being in the information/metadata, even if it replies completely on custom artifacts/logic); implementation still needs means to make decisions; initial reference implementation can be bare-bones, focused on parsing and logic tree, rather that physical aspects of deployment.
l. Require more details of classes of custom stuff; Rich to post an example, and it is suggested to customize the JRE example with custom stuff.
m. There is a suggestion to add an attribute indicating whether implementation has a choice to override built-in with custom; in any case, should document in spec.
Minutes for OASIS SDD TC 02 / 22 /2007 Meeting
AGENDA
11:05 AM ET Roll Call
1. Minutes approval Meeting Minutes from February 15 meeting.
2. Administrative (Julia) Also recent OASIS changes of note, for information (Brent):
- specification template changes, XML namespace conventions
- TC process changes: conformance statement, 3 statements of use
3. Ballot for next F2F meeting (need a motion)
4. Old/new business
- continue profile discussion?
- any other items
** 12:00 noon ET -- adjourn
MINUTES
We do not initially have a quorum. Quorum achieved at 11:10. (As always, attendance specifics can be seen on the TC web site in the meeting document – available from the Calendar view.)
1. Minutes for February 15, 2007 are approved without comment or objection.
2. Administrative
a. Please plan to complete your action items on schedule.
b. OASIS process changes:
i. There is a new OASIS specification template. Brent has forwarded it to Robert and Julia. He does not think we should update immediately since we are in the midst of major spec updates.
ii. There are also new OASIS TC process changes. OASIS formed a TC a while back to form a standard for OASIS specifications. One of the new requirements is that there must be a statement of conformance. Brent thinks our chapter on conformance that he has been working on will probably meet that requirement. Three statements of use are also required. These are written statements from TC members stating that they have successfully used the specification. For us, this probably means that we need at least 3 different members to create an implementation. We don’t think, at this point, that this has to be a shipping product. Clarification is needed on the details of the statement of use requirement.
3. Ballots now require a formal motion. So, we need a motion so that Brent can create the ballot. Motion made and seconded and passes without objection. Brent asks John if Dell can host the next meeting in Austin. This could be co-hosted with IBM Austin. e.g. part of the time at Dell, part of the time at IBM. John says it may be possible depending on the week chosen. The other choice is a similar arrangement using the SAS and IBM Raleigh facilities. Ballot will cover the last week in April and first two weeks in May.
4. We had some follow-on discussion of profiles. Issues discussed included:
a. representation of resource relationships in the dictionary; should these be represented via hierarchy or data within a flat list? Suggestion was made that we start with a flat list to avoid problems if it turns out a hierarchy can’t accurately represent relationships. We can always change to a hierarchy after we have a better idea of the types of relationships that will be modeled in the dictionary. It is noted that there are several relationship types implied by the contents of the SDD: hosted, component; consumes; and uses.
b. granularity of dictionary data; how “low” will it go? At the OS resource level we clearly need to go to the brand level (i.e. Windows, AIX). Will we go to the brand level for other types of resources? Maybe for something like a DB engine that hosts other types of resources. Maybe not for resources that do not host others.
c. Do we need to add something to the SDD to indicate the profiles required to support the SDD? Maybe. We’ll know better once we have a better idea of what profiles are.
d. How will the central dictionary be maintained? What will the infrastructure be? Who will support it? Who will pay for it?
e. Brent notes that our plans - for a specification with conformance levels as well as profiles to help determine interoperability - are consistent with the new OASIS conformance spec.
Minutes for OASIS SDD TC 02 / 15 /2007 Meeting
AGENDA
11:05 AM ET Roll Call
1. Minutes approval Meeting Minutes from February 8 meeting.
2. Administrative (Julia)
3. Review of company input about dictionary/profiles
** 12:00 noon ET -- adjourn
MINUTES
We have a quorum.
5. Minutes from 2/8/2007 are approved without comment or objection.
6. Processing of action items based on last week’s discussion has begun. Watch for update notifications for your action items later today.
7. Discussion of the dictionary and profiles. Brent has posted a presentation to guide the discussion. (Use the latest posted in the Documents view not the one linked to the most recent notification from Brent.) The discussion focused on clarifying the dictionary and profile concepts. I have attempted to note the highlights here.
a. An analogy: The dictionary is like a set of words to be used when creating profiles which are like sentences. In addition to the dictionary, we want to provide a set of starter/example/recommended profiles that spell out sentences that we believe are useful and make sense. Note that it will be possible to create profiles that do not make sense – like a sentence with bad grammar.
b. Obviously the dictionary will not be able to define all possible “words” in advance. When a profile creator needs a new “word”, should it be defined in the profile or defined as an extension of the dictionary? Assertion: New “words” should be extensions of the dictionary.
c. Someone suggested that a wikipedia model could be used to grow the dictionary. It was noted that version control is essential for interoperability so any wiki-based process would have to take that into account.
d. Where are relationships between “words” recorded? e.g. A hosts relationship between a database table resource and the database or a create association between a DDL artifact file and a database table. Assertion: only relationships that are always, irrevocably true should be recorded in the dictionary. Most relationships supported/understood by an implementation will be defined in the profiles.
Next steps: Brent and Randy will create a subset of a real example of the dictionary, profiles and SDDs that use the profiles. All are invited to do the same and bring their examples back to the group.
Minutes for OASIS SDD TC 02 / 08 /2007 Meeting
AGENDA
11:05 AM ET Roll Call
1. Minutes approval
A. Acceptance of Meeting Minutes from January 25 meeting.
B. Acceptance of Meeting Minutes from January 29 - Feb. 2 F2F meeting.
2. Administrative (Julia) [Note: we will have a review of selected new action items later in the agenda]
3. Specification updates (status, outlook, open items based on results of F2F meeting)
4. Review of selected new action items from F2F (objective is to plan for future calls -- proposals, profile discussion, etc.)
5. Other old/new business.
** 12:00 noon ET -- adjourn
MINUTES
We have a quorum.
1. Minutes approval
a. Minutes for January 25th are approved without objection.
b. Minutes for January 29th F2F are approved without objection.
2. Outstanding pre-January F2F action items.
a. Close 139 and open an action item for completing GGF alignment.
b. 162 looks like it was covered by the addition of file purpose types. Debra looks at use case 2.5.8 finds that this is really about overriding content in composed solutions. This is a CL2 issue and is still needed. Change date and add [CL2].
c. Close 173. Examples are posted in the Examples category on the TC web site document page.
d. Close 174
e. 175, 176 – change due date to 2 weeks from today – put on agenda.
f. 182 – need for this will be handled by profiles. Close.
g. 183 – split into CL1 and CL2 issues. Change CL1 date to next week.
h. 187, 201 – mark as CL2. Change date to middle of March
i. 189 – change date to end of February
j. 190 Change after discussion next week.
k. 191 – Close.
l. 193 – Close.
m. 197 – Close.
3. Spec updates are underway. Aggressive outlook is that a new version will be ready for TC review 2 weeks from today. We’ll checkpoint on this date next week.
4. Open proposals -
a. Profiles: We need everyone (one per company) to come back with ideas about what the dictionary and profile look like. There is a list on page 20 of the F2F meeting minutes of 10 or so aspects of profiles and dictionaries. Julia will open action items for each company to have these ideas prepared and posted in “Working Documents” for next week’s meeting.
b. Package types – change owner from Debra to Brent.
c. Human readable – Debra is still looking for feedback on the information grid she presented at the F2F. Who is planning to provide feedback? John will review. Feedback is needed in the next week and a half, and then Debra will put together a proposal for 3 weeks from today.
d. Version – Brent assigned to this. Debra suggests that this proposal wait until after the custom logic proposal.
e. Custom “Stuff” – Julia and Rich will work together on this with the goal of having a proposal 3 weeks from today.
f. Progress Indication – James and Chris will do this proposal.