Open Grid Forum: Data Format Description Language Working Group
OGF DFDL Working Group Call, May 12 -2010
Attendees
Steve Hanson (IBM)
Alan Powell (IBM)
Suman Kalia (IBM)
Stephanie Fetzer (IBM)
Tim Kimber(IBM)
Apologies
Steve Marting (Progeny)
Mike Beckerle (Oco)
1 Nils, defaults and unparsing
>From last calls minutes
- It was agreed that on unparsing a simple infoset item with no value will
be treated the same as missing. Steve H to check WMB behaviour. Need to
state in 4.1.2 Infoset that value is optional. "
Something to note. We say that for a simple element infoset item
[dataValue] has a value (either type compliant or nil) and for a complex
element infoset item [dataValue] has 'no value'. Also we say that for a
simple element infoset item [children] has 'no value' and for a complex
element infoset item [children] it may have a value (so it could also have
'no value'). The implication therefore is that an element infoset item is
deduced to be simple or complex based on the [dataValue] property. If we
let a simple item take 'no value' for [dataValue] as proposed, then there
is no reliable way to identify an element infoset item as simple or
complex.
Is that a problem in practice? After parsing, all simple items will have
[dataValue] set. For items created by some other means, and presumably
destined for unparsing, you could have an item with neither [dataValue]
nor [children] set. We could say (1) that an element infoset item is
known to be simple or complex only when it is processed by the unparser,
at which point the item is matched to the DFDL schema and its type
resolved. This is what broker's MRM parser does. Alternatively (2) we
introduce a simple/complex flag.
Note that I tried creating an MRM 'infoset' item for a string element but
just gave a name and no value. The MRM serializer looked at the element,
matched it to its model, found it was simple, and treated the value as
NULL which invoked null processing. The XML serializer also assumed NULL.
Personally I think we should adopt (1) but interpret 'no value' for a
simple item as nil and not missing.
Options for simple types:
1) Treat like missing from infoset. Means that an optional element will
not be output, and a required element will go through default value
processing.
2) Treat like [dataValue] is nil. Means nil processing is invoked if
nillable. This is what MRM does today.
3) Give an error - [dataValue] must always be set. It is up to the hosting
application to apply a semantic to a missing value from its own 'infoset'
and always set something when calling DFDL, if it wants to avoid such
errors.
After discussion general agreement that in the absence of any other reason
compatibility with existing products meant option 2. Steve H suggested we
get Mike's opinion.
Alan has been updating the Nils/Defaults section and raised the following
questions
A) Does nil processing apply to optional elements? In particular should a
nilValue of empty string apply to optional element.
Nil processing does apply to optional elements including a nilValue of
%#ES;
B) Does the nilValueDelimiterPolicy and emptyyValueDelimiterPolicy of a
child apply on parsing when the parent is missing. That is should there be
different rules for when the just child is missing than when the parent is
missing.
After discussion although it was felt that there could be formats that
required that support we weren't aware of any so V1 will say that the
child policy properties do apply.
2 OGF 29
Chicago, June 20-22
Suman volunteered to attend to present. Note the public comment period
ends 15 June.
3 Current Actions
see below
4 zoned numbers clarifications
The specification of the number pattern for zoned numbers is not clear.
A.
Clarification:
When dfdl:textNumberRep is ?zoned? only the pattern for positive numbers
is used. Only the following pattern characters may be used:
'+' MUST BE present at the beginning or end of the pattern to indicate
whether the leading or trailing digit carries the overpunched sign, if the
logical type is signed
'+' MAY BE present at the beginning or end of the pattern to indicate
whether the leading or trailing digit carries the overpunched sign, if the
logical type is unsigned
'V' MAY BE used to indicate the location of an implied decimal point
'0' indicates the number of required digits (including overpunched).
'#' indicates the number optional digits.
Agreed
B.
Also textNumberPolicy implies it applies to zoned, but doesn't state what
zoned behaviour it covers. I think it should be consistent with
binaryNumberCheckPolicy for packed, and control whether positive punched
data is accepted or rejected when parsing an unsigned type, and whether
unpunched data is accepted or rejected when parsing a signed type.
Agreed
C.
Finally, several textNumber properties only apply to standard or zoned. We
could distinguish these by name, which is what we have done elsewhere. So:
- textStandardBase
- textStandardGroupSeparator
- textStandardExponentCharacter
- textStandardInfinityRep
- textStandardNanRep
- textStandardZeroRep
- textZonedSignStyle
Stephanie noted that changes to properties caused rework of test cases so,
in general, should be resisted
Agreed to change
Meeting closed, 14:30
Next call Wednesday 19 May 2010 15:00 UK (10:00 ET)
Next action: 092
Actions raised at this meeting
No
Action
Current Actions:
No
Action
066
Investigate format for defining test cases
25/11:IBM to see if it is possible to publish its test case format.
04/12: no update
...
17/02: IBM is willing in principle to publish the test case format and
some of the test cases. May need some time to build a 'compliance suite'
24/03: No progress
03/03: Discussions have been taking place on the subset of tests that will
be provided.
10/03: work is progressing
17/03: work is progressing
31/03: work is progressing
14/04: And XML test case format has been defined and is being tested.
21/04. Schema for TDML defined. Need to define how this and the test cases
will be made public
05/05: Work still progressing
12/05: Work still progressing
084
Check behaviour of dfdl:inputValueCalc and outputValueCalc.
14/04: no prorgess
21/04: no progress
05/05: no progress
085
ALL: publicize Public comments phase to ensure a good review..
14/04: see minutes
21/04: Press release, OMG and other standards bodies.
05/05: Alan and Steve H have contacted other standards bodies. Will ask
them to add comments on spec
15/05: still no public comments
086
AP: Nils and Defaults during unparsing - update table
31/03: TK to documetn use cases for parsing
14/04: Investigate new property to control empty string behaviour.
21/04: After investigation a new property is not required. New rules
developed and tables updated.
Need examples of complexTypes to confirm tables apply.
Review Nils, defaulting spec section.
05/05: Discussed defaulting complex elements. Tables updated but need to
add terminator.
SH; to confirm WMD behaviour when infoset item has no value on unparsing
Need to describe defaulting choices.
15/05: More discussion. Alan updating sections
088
define semantics of choiceKind 'fixedLength'
31/03: TK to provide definition of calculable length.
Investigate PL/I varchars and Cobol occurs dependingon.
14/04Tim had distributed a document starting the definition of calculable
length for the longest choice member.
Alan had done some investigation of COBOL occurs depending on and when
used in the working section of a program then the maximum storage was
reserved but when used in the linkage section the dependent number was
uses. We need to understand how the WMB COBOL importer deals with ODO.
21/04: Need to define 'calculable length' and WMB importer ODO behaviour.
05/05: TK: Still need definition of calculable length.
SKK: WMB COBOL imported behaviour with ODO
15/05: Suman sent an expmle of an imported Cobol ODo which suggested that
the maximum space was reserved. He will extend the example.
091
The IBM team had questioned whether timezones should be just UTC+/- offset
or should GMT and other short time zones be allowed. Alan to investigate
15/05: Alan to talk to Ruth.
Closed actions
No
Action
Work items:
No
Item
target version
status
005
Improvements on property descriptions
not started
012
Reordering the properties discussion: move representation earlier, improve
flow of topics
not started
036
Update dfdl schema with change properties
ongoing
042
Mapping of the DFDL infoset to XDM
none
not required for V1 specification
070
Write DFDL primer
071
Write test cases.
083
Implement RFC2116
097
Remove functions that returns duration
098
occursCountKind is parsing only
099
nilValue and OccuresStopValue cannot have an expression.
On unparsing only outputValueCalc can have a forward reference.
100
Need to state in 4.1.2 Infoset that value is optional.
101
When dfdl:textNumberRep is ?zoned? only the pattern for positive numbers
is used. Only the following pattern characters may be used:
'+' MUST BE present at the beginning or end of the pattern to indicate
whether the leading or trailing digit carries the overpunched sign, if the
logical type is signed
'+' MAY BE present at the beginning or end of the pattern to indicate
whether the leading or trailing digit carries the overpunched sign, if the
logical type is unsigned
'V' MAY BE used to indicate the location of an implied decimal point
'0' indicates the number of required digits (including overpunched).
'#' indicates the number optional digits.
102
Also textNumberPolicy implies it applies to zoned, but doesn't state what
zoned behaviour it covers. I think it should be consistent with
binaryNumberCheckPolicy for packed, and control whether positive punched
data is accepted or rejected when parsing an unsigned type, and whether
unpunched data is accepted or rejected when parsing a signed type.
103
- textStandardBase
- textStandardGroupSeparator
- textStandardExponentCharacter
- textStandardInfinityRep
- textStandardNanRep
- textStandardZeroRep
- textZonedSignStyle
104
4 Unsigned decimal
Proposed solution:
- Allow xs:nonNegativeInteger which enables unsigned unbounded integers to
be modelled. The problem to solve is then just for xs:decimal.
- Call the new property dfdl:decimalSigned. It only applies to xs:decimal
or user defined restrictions thereof. It applies to all physical
decimals, as its name implies (not just zoned or packed).
Regards
Alan Powell
Development - MQSeries, Message Broker, ESB
IBM Software Group, Application and Integration Middleware Software
-------------------------------------------------------------------------------------------------------------------------------------------
IBM
MP211, Hursley Park
Hursley, SO21 2JN
United Kingdom
Phone: +44-1962-815073
e-mail: alan_powell(a)uk.ibm.com
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
