Rather than overloading "implicit" length to mean this, I think we need an enum-valued property for choices.

Today there is this choiceUnresolvableWhenParsing boolean. I think this shoudl become choiceType="unresolvableWhenParsing".

Then we can have an enum value for the case where all arms of the choice are assumed to be padded to max length of the alternative.

So I think we have 3 enum values:

variableLength
fixedLength
unresolvableWhenParsing

variableLength means the storage taken depends on the alternative selected, Fixed length means shorter alternatives are padded to the length of the longest. unresolvable means what it means now, which is that you can't tell which variant it is (which also implies the storage required is fixed length).

...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

Steve Hanson/UK/IBM@IBMGB

11/19/2007 07:13 AM

To	Mike Beckerle/Worcester/IBM@IBMUS
cc	dfdl-wg@ogf.org
Subject	Re: [DFDL-WG] Grammar as adapted to infosetLink

Or allow xs:choice to carry dfdl:length & dfdl:lengthKind. Then lengthKind="implicit" could mean the length is the longest of the modelled alternatives.

Regards, Steve

Steve Hanson
WebSphere Message Brokers
Hursley, UK
Internet: smh@uk.ibm.com
Phone (+44)/(0) 1962-815848

Mike Beckerle/Worcester/IBM@IBMUS

16/11/2007 19:29

To	Steve Hanson/UK/IBM@IBMGB
cc	dfdl-wg@ogf.org
Subject	Re: [DFDL-WG] Grammar as adapted to infosetLink

There are of course alternatives.

E.g., a hidden or non-hidden field which is a byte array or hexBinary to absorb the padding bytes at the ends of the shorter variants. Or one could put trailing skip bytes on the last element or group of each of the shorter variants.

One can also encapsulate the whole choice in a sequence whose length is the fixed maximum length. This is my personal favorite, since it uses the FinalUnused that is already in the grammar.

E.g., suppose the length of the longest variant is 258 bytes.

<sequence dfdl:lengthKind="explicit" dfdl:length="258" dfdl:lengthUnits="bytes" dfdl:applies="hereOnly">
<choice>
.....
</choice>
</sequence>

The only thing I don't like about this is I don't have a way to express the length other than to hard-code the constant 258, but since tooling would typically put this in based on Cobol descriptors or other information I don't mind so much.

...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

Steve Hanson/UK/IBM@IBMGB

11/16/2007 11:24 AM

To	Mike Beckerle/Worcester/IBM
cc	dfdl-wg@ogf.org
Subject	Re: [DFDL-WG] Grammar as adapted to infosetLink

Hi Mike

One extra thought post-call, caused by thinking about the FinalUnused part of a sequence. Some fixed length choices have a rule that each branch of the choice must be of equal length. This results in unused bytes for branches shorter than the maximum. In a bitstream instance, a user will see these bytes. Perhaps the grammar should acknowledge their existence, even if there are no DFDL properties to express the bytes (like FinalUnused)?

Regards, Steve

Steve Hanson
WebSphere Message Brokers
Hursley, UK
Internet: smh@uk.ibm.com
Phone (+44)/(0) 1962-815848

Mike Beckerle <beckerle@us.ibm.com>
Sent by: dfdl-wg-bounces@ogf.org

10/11/2007 00:16

To	dfdl-wg@ogf.org
cc
Subject	[DFDL-WG] Grammar as adapted to infoset

I took a crack at this. I ripped it out of the main document for faster review/turnaround.

I think this matches the latest infoset email diagram.

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
-- dfdl-wg mailing list dfdl-wg@ogf.org http://www.ogf.org/mailman/listinfo/dfdl-wg