Due to an issue with CutePDF truncating
text, here is the agenda for today.
1. Spec issues from IBM
As implementation has progressed, a
number of issues have been spotted in the DFDL 1.0 spec by the IBM implementation
team.
List so far for resolution.
6.3. Bullet for logical value. State
that the string must obey the lexical representation of the type.
6.3.1. Explicit statement of white space
rules is missing. Just white space for a DFDL string literal property is
an error. When a property can be a list of string literals, white space
between values is collapsed.
7.7. What is the scope of a variable
with respect to included/imported xsds? Can I refer from xsd A to a variable
defined in parent xsd B?
9.2. Grammar. Does not allow for a prefix
length itself having a prefixed length, so does not match property description
(ie, the change to permit this is draft 045 omitted to update the grammar).
13. Be explicit on the DFDL literals
allowed for text properties. See separate e-mail.
13.6. When textNumberRep is zoned, property
description should state that base is assumed to be 10.
13.6. State that dfdl:ignoreCase not
honoured by textStandardExponentCharacter, textStandardInfinityRep,
textStandardNanRep as ICU does not offer that flexibility.
13.6 Allow multiple characters for textStandardExponentCharacter
to handle reps like 1.23x10^4
as ICU allows that. Note that property name will have to change, eg, to
textStandardExponent..
13.6. Change meaning of textNumberCheckPolicy
'lax' to align more closely with ICU:
"If ‘lax' and dfdl:textNumberRep is
'standard' then grouping separators are ignored, leading and trailing whitespace
is ignored, leading zeros are ignored."
13.6. Disallow the use of empty string
for textStandardExponentCharacter. State property must be set if the pattern
contains an 'E'
13.6. Disallow the use of empty string
for textStandardDecimalSeparator, State property must be set if the pattern
contains a '.'
13.6 Disallow the use of empty string
for textStandardGroupingSeparator, State property must be set if the pattern
contains a ','
13.6. Allow decimal
separator to be a List of DFDL String Literals or a DFDL expression. This
allows modelling of EDIFACT where
user can choose decimal separator in ISA header but '.'
is also allowed.
13.6. When textNumberPadCharacter is
'0' which it commonly is, a value of say '00000' will get trimmed to the
empty string, whereas the intent would surely be to trim to '0'. Need a
rule for this.
13.6.Should textStandardDecimalSeparator
ignored when the logical type is not decimal/float/double?
13.6.1.1. Erroneously mentions sigDigits
in reference to the BNF for textNumberPattern.
13.6.1.1. Formatting. Uses terms 'minimum/maximum
integer/fraction digits' but does not define them.
13.9. Should state that textBooleanTrue/FalseRep
properties are used after trimming when parsing, and before padding when
unparsing. If lengthKind is explicit or implicit and either textPadKind
or textTrimKind is none then the properties must have the same length else
schema definition error.
13.16. Nil literal character. When
representation is binary, the encoding property is not used, which implies
that a nil value must only contain DFDL hex literals (%rXX;)
13.16. Clarify nilKind 'literalCharacter'
and 'literalValue' and whether nil test is applied before or after trimming.
15. Using a discriminator to resolve
each branch of a choice with a very large number of branches has a performance
impact as each branch must be evaluated in turn. In cases where the name
of the branch element is known or derivable in advance (ie, each discriminator
expression is of the form {..\tag eq "elem1"} or similar) then
consider adding a property to dfdl:choice that provides an expression that
returns a QName, being the name of the branch element. Note: IBM MRM has
something like this.
16.1. Revise entire section.
16.2. State it is a processing error
if the stop value is missing when parsing.
2. Spec issues from Mike Beckerle
A number of issues have been spotted
by Mike. Some have been resolved by action 139 below, but others remain.
See separate e-mail.
Regards
Steve Hanson
Architect, Data Format Description Language (DFDL)
Co-Chair, OGF
DFDL Working Group
IBM SWG, Hursley, UK
smh@uk.ibm.com
tel:+44-1962-815848
----- Forwarded by Steve
Hanson/UK/IBM on 28/06/2011 14:46 -----
From:
| "Mike Beckerle" <mbeckerle.dfdl@gmail.com>
|
To:
| Steve Hanson/UK/IBM@IBMGB
|
Date:
| 28/06/2011 13:50
|
Subject:
| RE: [DFDL-WG] OGF DFDL WG Call Agenda
2011-06-28 |
Steve,
When I download this agenda
PDF file, it’s all cut off on the right edge of the page, making it hard
to figure out what some of these line items are about.
Something is up with the
way you converted this teamroom page to a pdf.
If this was paper, I’d suspect
the metric vs. US paper size issue, but this is all bits…. J
…mikeb
From: dfdl-wg-bounces@ogf.org [mailto:dfdl-wg-bounces@ogf.org]
On Behalf Of Steve Hanson
Sent: Monday, June 27, 2011 12:18 PM
To: dfdl-wg@ogf.org
Subject: [DFDL-WG] OGF DFDL WG Call Agenda 2011-06-28
Please find agenda for the above call on GridForge at:
http://forge.gridforum.org/sf/docman/do/downloadDocument/projects.dfdl-wg/docman.root.current_0.calls/doc16276/2
Regards
Steve Hanson
Architect, Data Format Description Language (DFDL)
Co-Chair, OGF
DFDL Working Group
IBM SWG, Hursley, UK
smh@uk.ibm.com
tel:+44-1962-815848
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
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