Here's my
notes:
DFDL WG Semi-Weekly
Call 2005-04-13
Attendees
Jim Meyers
Tara
Talbot
Steve Hanson
Mike Beckerle
Karen Loughran
Agenda
1) Scope of the v1.0 specification
(not to be confused with "scoping of annotations")
2) Discussion of scoping
of annotations based on Ascential posting.
3) Planning of agenda for upcomign
F2F meeting on May 11.12, 13, at Ascential in Westborough,
MA
Suggested topics so far:
- Review of TD Model and
extension to cover DFDL properties
- Examples from the Ascl prototype - go
through several line by line to see what we do and don't like about them-
F2F Who's coming:
Mike Beckerle
Jim
Meyers
Tara Talbot
Suman Kalia
Steve Hanson
Who's not
coming:
Martin Westhead
Karen Loughran
Scope of v1.0 Spec.
-
Steve Hanson - don't
bite off too much for v1.0
- QSAM formats with data split over other
pieces.
- Encrypted sections and how we handle those.
- DFDL Validation
i.e., check that the binary data is valid according to the DFDL Schema.
(It's alot of work to do a validating parser. EDIFACT, X12 and such
require more complex validation than just. Things like field a if present means
field b should NOT be present.)
Jim Meyers -
multi-layer stuff is what supports those things.
Scoping rules
discussion -
Defer discussion -
Agenda for F2F -
- model and parser
behavior aspects: rules for default values, null handling, how to figure out how
to figure ot the length precedence rules, etc.
- open content - tagged, but
some of the tags are unknown. HL7 people put extra fields into a group. Happens
with every format though, not just tags. Polymorphic chunks. xsd:all - unordered
data.
- PNNL project and Jaxme implementation - here's what's on the ground.
Focus should be on how that's changing the language.
- Scoping of
annotations
- extensibility mechanisms (low prio)
- multi-dimensional
arrays (low prio)
- reader/writier/filter/function - topic
- don't bother with
a deep dive on any implementation.