This is where Suman and I left it on
the array prefix issue.
The problem is where do you stop with
creating special "occurs" versions. You can in principle have
an "occurs" variant of almost every property.
...mikeb
Mike Beckerle
STSM, Architect, Scalable Computing
IBM Software Group
Information Platform and Solutions
Westborough, MA 01581
direct: voice and FAX 508-599-7148
assistant: Pam Riordan
priordan@us.ibm.com
508-599-7046
----- Forwarded by Mike
Beckerle/Worcester/IBM on 11/20/2007 10:02 AM -----
| Suman Kalia/Toronto/IBM@IBMCA
11/16/2007 11:53 AM
|
|
To
| Mike Beckerle/Worcester/IBM@IBMUS
|
|
cc
|
|
|
Subject
| Fw: DFDL array prefix |
|
Mike - The proposed grammar productions
seem correct to me..
I see your view point and your concern
for bifurcation of properties and I do agree that it introduces level of
redundancy. I guess the choice boils down to having extra level of redundancy
(dummy sequence) in the logical model or having redundancy in DFDL properties
and each alternative can be justified and supported with good arguments.
>From consumability perspective having clean and less/no extra level of
redundancy in the logical model is preferred because it would be
at the user face all the time while the DFDL properties pertaining to occurrences
will be rare and only for specific use cases. From pure DFDL spec
perspective, having no redundancy in DFDL properties is desirable.
May be we can go with the conservative
approach as you mentioned ie not add the extra production in
the grammar for Array for now but add an appendix in the document listing
the alternative for overcoming redundancy in the logical model and how
the revised grammar would look like. We can get more feedback when the
document is reviewed by IBMers (architects, implementers, UCD) and folks
in the standards community.
Suman Kalia
IBM Toronto Lab
WebSphere Business Integration Application Connectivity Tools
Tel : 905-413-3923 T/L 969-3923
Fax : 905-413-4850 T/L 969-4850
Internet ID : kalia@ca.ibm.com
----- Forwarded by Suman
Kalia/Toronto/IBM on 11/16/2007 10:34 AM -----
| Mike Beckerle/Worcester/IBM@IBMUS
11/16/2007 10:17 AM
|
|
To
| Suman Kalia/Toronto/IBM@IBMCA
|
|
cc
|
|
|
Subject
| DFDL array prefix |
|
Here's the proposal I think:
Array = ArrayPrefix ArrayContent
ArrayContent = [ Element [ OccursSeparator
Element ]* [ OccursSeparator StopValue] ]
StopValue = SimpleElement
I'd suggest the keyword occursLeadingSkipBytes
is the property which controls the ArrayPrefix content.
My argument against this is a slippery-slope
argument. This starts us down the path of having a bifurcation of
many properties into an occurs and non-occurs flavor. E.g., once you add
the above, then by analogy why not have these additional keywords which
apply to the array as a whole, not the elements of it:
occursAlignment
occursAlignmentUnits
occursInitiator
occursTerminator
occursLength
occursLengthKind
occursLengthUnits
All of this is avoided by simply requiring
an extra sequence to provide these same format characteristics using the
ordinary non-occurs keywords, and that's very preferable to me to this
proliferation of occurs and non-occurs variants.
...mikeb
Mike Beckerle
STSM, Architect, Scalable Computing
IBM Software Group
Information Platform and Solutions
Westborough, MA 01581
direct: voice and FAX 508-599-7148
assistant: Pam Riordan
priordan@us.ibm.com
508-599-7046