John;

My first reaction would have been to say definitely yes.  I've always
found that behavior an unintuitive gotcha for which I could get burnt
without warning.  I've learned to be more careful, but I still get
zapped every once in a while.  Teams might be more impacted.  (Why does
this work here and not there?  Why did Sam do that?)  Sometimes the
workaround is unclean.

On further thought, however, it might create real headaches for the
unaware who are switching between structure and non-structure methods.
The cure could be worst than the disease.  

(What happens if you define a 407E break field past the end of the
record?  Am very busy right now, otherwise I'd try it first.)

Have you considered a keyword to switch this behavior on and off, either
for the entire specs stage, or from a given point forward?  I think this
would be the safest.  It would preserve existing functionality and
compatibility while providing for more intuitive behavior and a simpler
workaround.

Just my 2 cents.  Accept, reject or modify as you see fit.

- Hobart 

-----Original Message-----
From: CMSTSO Pipelines Discussion List
[mailto:[EMAIL PROTECTED] On Behalf Of John Hartmann
Sent: Thursday, October 09, 2008 9:53 AM
To: [email protected]
Subject: [CMS-PIPELINES] Null input fields with structured output.

The SPEC reference states:

"A data field is suppressed if it refers to an input range that is not
present  in  the  record  and the output placement does not specify an
explicit length for the output field; the item is then ignored."

When the output placement is a member of a structure it will have a
length and thus such a specification item will never be ignored.  But
is this desirable?

     j.

Reply via email to