[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: FW: [provision] open xsd issues.
*12. ObjectClassRefType/SchemaEntityRefType must contain
*exactly one of* "objectClassName" (String) or "elementName"
(QName).
(It would be weird to have both and useless to have neither.)
20041201 - Sent email: "Draft 9 Item 12:
SchemaEntityRefType#objectClassName and #elementName."
20041206 - Gerry Woods suggests that a single URI-valued or
string-valued attribute
(e.g., 'entityName') could replace both objectClassName
and elementName.
*15. Remove (optional) requestID from SpmlRequestType
and add (required) requestID to CancelRequestType
and add (required) requestID to StatusRequestType
DISPUTED.
20041201 - email: "Issue 15, request IDs for synchronous requests."
20041206 - Leave 'requestID' as-is and add 'asyncOperationID' to cancel
and status?
Draft 10 - Still need 'asyncRequestID' in spml:ResponseType.
*16. Don't need CancelResultsType; ResultCode and ErrorCode should
suffice.
DISPUTED.
20041201 - email: "Issue 16, cancel results..."
*17. Replace "statusReturns" with optional boolean attribute
"returnOutput".
DISPUTED.
20041201 - email: "Issue 17, statusReturns..."
20051205 - Rename ResultCode to StatusCode and rename attribute result
to status
and let the optional boolean attribute be named
"returnResults".
----------------------
*25. The "active" attribute of ActiveResponse should be optional.
(In case of error, any value of "active" could be misleading.)
OK? - Optional, but having an element seems like overkill. Suggest
attribute.
20041130 - sent email: "Draft 9: ActiveResponseType#active."
25. SelectionType#path and #namespace.
Draft 9 adds namespace mapping but removes required <namespace>
element.
I thought we needed this to know the query language of the "path"
value.
20041201 - Sent email: "Draft 9: SelectionType#path and #namespace."
20041206 - Gerry says that we must ask Jeff.
26. PSOType#psoId and #data. (FIXED in Draft 10).
<psoId> and <data> are *optional* sub-elements of PSOType.
I think these should be required.
20041201 - Sent email: "Draft 9: PSOType#psoId and #data."
20041206 - Rami made the same point during face-to-face review.
Draft 10 - Fixed.
27. Rename ChangeType to ModificationModeType.
20041201 - Sent email: "Draft 9: Rename ChangeType to
ModificationModeType."
20041206 - Accepted in face-to-face review.
28. Rename ExecutionType to ExecutionModeType.
20041201 - Sent email: "Draft 9: Rename ExecutionType to
ExecutionModeType."
20041206 - Accepted in face-to-face review.
29. Camel-case "objectclassName" and "elementName" in
SchemaEntityRefType.
20041206 - Accepted in face-to-face review.
30. Element ref won't work.
20041201 - send email "element ref won't work".
StatusResponseType must nest 1+ response derived from
SpmlResponseType.
BatchRequestType must nest 1+ requests derived from
BatchableRequestType.
BatchResponseType must nest 0+ responses derived from
SpmlResponseType.
Options:
1) CHOICE of AddRequestType|ModifyRequestType|DeleteRequestType.
pro: xsd validation
con: not extensible
2) WRAPPER: similar to <data> and <capabilityData>
pro: open; xsd validation as to minOccurs, maxOccurs; know
where to look
con: cannot validate "payload" of wrapper
3) XSI-TYPE: XSD specifies base type, instance specifies derived
type.
pro: intuitive; easy to specify; recommended in XML Schema
Primer
con: fancy?
4) OPEN CONTENT: 1-N requests appear as open content of
batchRequest.
pro: open
con: no xsd validation whatsoever; no indication as to
location;
specified behavior left as open content;
inconsistent with <data> and <capabilityData>
5) CHOICE+OPEN CONTENT: Choice for core and open content for
capabilities
31. AddRequestType#container won't work. (FIXED in Draft 10)
20041202 sent email "AddRequestType#container won't work."
Options (I like 2 best):
1) XSI-TYPE: XSD specifies base type, instance specifies derived
type.
2) CHOICE of IdentifierType | PSOIdentifierType
3) Use one type for both (with optional containerId/targetId)
20041206 - Face-to-face accepted choice of IdentifierType |
PSOIdentifierType.
Draft 10 - Fixed.
--------------------------
>From Face-to-Face 20041206
--------------------------
32. Allow 'status' operation to serve as 'statusAll'
- 'asyncOperationId' becomes optional
- if no 'asyncOperationId', return all
- <nestedResponse minOccurs="0" maxOccurs="unbounded">
33. Requirement 4.4: Multiple Error Messages. (FIXED in Draft 10)
- leave ErrorCode as-is
- add <errorMessage minOccurs="0" maxOccurs="unbounded">
Draft 10 - Fixed.
32. Rename SchemaEntityRefType#supports to #supportedSchemaEntity
33. TargetType#targetID should be minOccurs="1" (default).
34. TargetType#capability should be minOccurs="0".
35. ListTargetsResponseType#target should be minOccurs="0" (in case
of error).
36. Rename CapabilityType#identifier to #namespace (more specific).
37. Add TargetType#capabilities element to wrap <capability>
elements.
38. Requestor must "program-by-exception" wrt containment. Is this
okay?
- nothing tells requestor whether target supports containment.
- nothing tells requestor whether target schema entity supports
containment.
- Do we need a Containment capability to advertise either or
both of these?
39. Should a single <modification> contain both <component> and
<capabilityData>?
40. What does an instance of {SelectionType} contain in the SMPL 1.0
(DSML) Profile?
41. Wouldn't it be easier to say that modifyResponse returns entire
<pso>?
What provider wants to calculate all the differences?
What requestor wants to process all the differences?
- optional <psoId> (if changed)
- boolean option returnData (or just always return entire pso)?
- boolean option returnCapabilityData (since this could be
huge)?
42. Should a <deleteRequest> contain <capabilityData>?
-----Original Message-----
From: Gary P Cole [mailto:Gary.P.Cole@Sun.COM]
Sent: Tuesday, December 07, 2004 6:19 AM
To: PSTC
Subject: [provision] open xsd issues.
Attached is a list of issues that were open against draft 9 of the XSD
(updated as of 20041206). Changes (from last night's version):
- Summarized the proposals for issue #30.
- Draft 10 fixes PSOType#data and #id issue #26.
- Draft 10 fixes AddRequestType#container issue#31.
- Draft 10 fixes Multiple Error Messages issue #33.
gpc
*12. ObjectClassRefType/SchemaEntityRefType must contain
*exactly one of* "objectClassName" (String) or "elementName" (QName).
(It would be weird to have both and useless to have neither.)
20041201 - Sent email: "Draft 9 Item 12: SchemaEntityRefType#objectClassName and #elementName."
20041206 - Gerry Woods suggests that a single URI-valued or string-valued attribute
(e.g., 'entityName') could replace both objectClassName and elementName.
*15. Remove (optional) requestID from SpmlRequestType
and add (required) requestID to CancelRequestType
and add (required) requestID to StatusRequestType
DISPUTED.
20041201 - email: "Issue 15, request IDs for synchronous requests."
20041206 - Leave 'requestID' as-is and add 'asyncOperationID' to cancel and status?
Draft 10 - Still need 'asyncRequestID' in spml:ResponseType.
*16. Don't need CancelResultsType; ResultCode and ErrorCode should suffice.
DISPUTED.
20041201 - email: "Issue 16, cancel results..."
*17. Replace "statusReturns" with optional boolean attribute "returnOutput".
DISPUTED.
20041201 - email: "Issue 17, statusReturns..."
20051205 - Rename ResultCode to StatusCode and rename attribute result to status
and let the optional boolean attribute be named "returnResults".
----------------------
*25. The "active" attribute of ActiveResponse should be optional.
(In case of error, any value of "active" could be misleading.)
OK? - Optional, but having an element seems like overkill. Suggest attribute.
20041130 - sent email: "Draft 9: ActiveResponseType#active."
25. SelectionType#path and #namespace.
Draft 9 adds namespace mapping but removes required <namespace> element.
I thought we needed this to know the query language of the "path" value.
20041201 - Sent email: "Draft 9: SelectionType#path and #namespace."
20041206 - Gerry says that we must ask Jeff.
26. PSOType#psoId and #data. (FIXED in Draft 10).
<psoId> and <data> are *optional* sub-elements of PSOType.
I think these should be required.
20041201 - Sent email: "Draft 9: PSOType#psoId and #data."
20041206 - Rami made the same point during face-to-face review.
Draft 10 - Fixed.
27. Rename ChangeType to ModificationModeType.
20041201 - Sent email: "Draft 9: Rename ChangeType to ModificationModeType."
20041206 - Accepted in face-to-face review.
28. Rename ExecutionType to ExecutionModeType.
20041201 - Sent email: "Draft 9: Rename ExecutionType to ExecutionModeType."
20041206 - Accepted in face-to-face review.
29. Camel-case "objectclassName" and "elementName" in SchemaEntityRefType.
20041206 - Accepted in face-to-face review.
30. Element ref won't work.
20041201 - send email "element ref won't work".
StatusResponseType must nest 1+ response derived from SpmlResponseType.
BatchRequestType must nest 1+ requests derived from BatchableRequestType.
BatchResponseType must nest 0+ responses derived from SpmlResponseType.
Options:
1) CHOICE of AddRequestType|ModifyRequestType|DeleteRequestType.
pro: xsd validation
con: not extensible
2) WRAPPER: similar to <data> and <capabilityData>
pro: open; xsd validation as to minOccurs, maxOccurs; know where to look
con: cannot validate "payload" of wrapper
3) XSI-TYPE: XSD specifies base type, instance specifies derived type.
pro: intuitive; easy to specify; recommended in XML Schema Primer
con: fancy?
4) OPEN CONTENT: 1-N requests appear as open content of batchRequest.
pro: open
con: no xsd validation whatsoever; no indication as to location;
specified behavior left as open content;
inconsistent with <data> and <capabilityData>
5) CHOICE+OPEN CONTENT: Choice for core and open content for capabilities
31. AddRequestType#container won't work. (FIXED in Draft 10)
20041202 sent email "AddRequestType#container won't work."
Options (I like 2 best):
1) XSI-TYPE: XSD specifies base type, instance specifies derived type.
2) CHOICE of IdentifierType | PSOIdentifierType
3) Use one type for both (with optional containerId/targetId)
20041206 - Face-to-face accepted choice of IdentifierType | PSOIdentifierType.
Draft 10 - Fixed.
--------------------------
>From Face-to-Face 20041206
--------------------------
32. Allow 'status' operation to serve as 'statusAll'
- 'asyncOperationId' becomes optional
- if no 'asyncOperationId', return all
- <nestedResponse minOccurs="0" maxOccurs="unbounded">
33. Requirement 4.4: Multiple Error Messages. (FIXED in Draft 10)
- leave ErrorCode as-is
- add <errorMessage minOccurs="0" maxOccurs="unbounded">
Draft 10 - Fixed.
32. Rename SchemaEntityRefType#supports to #supportedSchemaEntity
33. TargetType#targetID should be minOccurs="1" (default).
34. TargetType#capability should be minOccurs="0".
35. ListTargetsResponseType#target should be minOccurs="0" (in case of error).
36. Rename CapabilityType#identifier to #namespace (more specific).
37. Add TargetType#capabilities element to wrap <capability> elements.
38. Requestor must "program-by-exception" wrt containment. Is this okay?
- nothing tells requestor whether target supports containment.
- nothing tells requestor whether target schema entity supports containment.
- Do we need a Containment capability to advertise either or both of these?
39. Should a single <modification> contain both <component> and <capabilityData>?
40. What does an instance of {SelectionType} contain in the SMPL 1.0 (DSML) Profile?
41. Wouldn't it be easier to say that modifyResponse returns entire <pso>?
What provider wants to calculate all the differences?
What requestor wants to process all the differences?
- optional <psoId> (if changed)
- boolean option returnData (or just always return entire pso)?
- boolean option returnCapabilityData (since this could be huge)?
42. Should a <deleteRequest> contain <capabilityData>?
To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/provision/members/leave_workgroup.php.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]