Open Grid Forum: Data Format Description
Language Working Group
Working Group Conference Call
14:00 GMT, 16 Nov 2007
Attendees
Mike Beckerle (IBM)
Geoff Judd (IBM)
Steve Hanson (IBM)
Suman Kalia (IBM)
Ian Parkinson (IBM)
Alan Powell (IBM)
1. Introduction
Mike called the meeting to discuss the
revised Data Grammar portion of the DFDL specification, which he had previously
circulated. This was discussed on the regular, weekly, conference call
on 14th November, in Mike's absence. Mike will not be able to attend the
next scheduled conference call.
2. Review of DFDL Data Grammar
The meeting discussed the following
points:
- As noted in the previous meeting,
both the ComplexContent and ComplexElement productions include a "Prefix"
element. Mike stated that this redundancy does exist in DFDL, but it sometimes
necessary, e.g. for embedded groups. Steve suggested that, if this did
need to be changed, we would do so with restrictions rather than a change
to the productions.
- Suman suggested that the present
grammar specification does not satisfactorily cover the case where the
start of an array needs to be aligned, with individual array elements automatically
being aligned. Mike suggested that this can be dealt with, in a DFDL schema,
by encapsulating the array in an element or sequence.This would require
the author to build additional encapsulation around the array, which Suman
felt would be unnatural, and suggested it would be more natural to introduce
a property to explicitly deal with this situation. Mike observed that the
XML specification treats arrays as simply adjacent elements, so we should
minimize the number of properties which affect an entire array rather than
each individual element. Mike and Suman will collaborate on a concrete
proposal to address this.
- Mike expressed discomfort with
the use of "binary" and "opaque" to describe raw data,
as these words ,ay be ambiguous. "opaque" in particular may refer
to unparsed data, or to anyType elements. Suggested alternatives included
"octets", "bits", "raw" and "blob".
- The SequenceContent production
currently contains a "finalUnused" element. Mike wondered whether
this would best be moved to the Sequence production. Suman felt that there
would then be issues with prefixes and suffixes. The meeting saw no good
reason to change this.
- Precedence order needs to be
defined. Steve and Geoff will attempt to define a static order.
- Regarding the query raised at
the previous meeting concerning initial prefix Separators and final postfix
Separators, Mike confirmed that these had been removed following comments
in a previous draft of the document. They had originally been intended
to support certain properties which could be now be defined using other
features of the grammar.
Meeting closed 15:00 GMT
Ian Parkinson
WebSphere ESB Development
Mail Point 211, Hursley Park, Hursley, Winchester, SO21 2JN, UK
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