Re: issue #144 proposal - array metadata in SOAP Encoding (long)

That was indeed the idea, but just an idea.  If everyone who has been 
involved in using and implementing the encodings thinks it's a good one, 
fine.  Otherwise, I certainly wouldn't push it.  It just seems to me to 
make the models for explicit and implicit positions more consistent.  Is 
it not in general allowed to mix the two?  If so, then how would you deal 
with:

        <A Position='3'> ... </A>
        <A Position='4'> ... </A>
        <A Position='2'> ... </A>
        <A> ... </A>  <! -- dup of position 3 -->
        <A> ... </A>  <! -- dup of position 4 -->

Are the last two duplicates?  I think that by requiring everything to be 
ordered, you avoid having to deal with rules for strange situations such 
as this.  You still have to deal with:

        <A Position='2'> ... </A>
        <A> ... </A>
        <A> ... </A>
        <A Position='3'> ... </A> <! -- dup of position 3 -->

Which I presume to be illegal, right?

------------------------------------------------------------------------
Noah Mendelsohn                                    Voice: 1-617-693-4036
Lotus Development Corp.                            Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------------







Jacek Kopecky <jacek@systinet.com>
Sent by: xml-dist-app-request@w3.org
10/31/01 10:47 AM

 
        To:     <Noah_Mendelsohn@lotus.com>
        cc:     <rsalz@zolera.com>, <xml-dist-app@w3.org>
        Subject:        Re: issue #144 proposal - array metadata in SOAP Encoding (long)

 Noah,
 would you also like all the positions to be in order?
 I don't think adding these constraints would be a bad idea as I
feel there may be cases where knowing the elements are in order
could help.
 It would be consistent, too, but I don't think the text without
these constraints is inconsistent.
 Anyway, I think I can support adding these two rules as phrased
below (and with possible editorial changes of course):
 "The presence of the enc:offset attribute indicates the
partially transmitted array contains no member on position below
the offset value."
 and
 "The members in a partially transmitted array must appear in
order, i.e. the rightmost index is changing most rapidly."

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)
                   http://www.systinet.com/



On Wed, 31 Oct 2001 Noah_Mendelsohn@lotus.com wrote:

 > If both offset and positions are specified, would it be better or worse 
to
 > require that no position preceeds the offset?  That would allow you to
 > have implementations where there is, in general, a current offset 
starting
 > at 0 in all dimensions.  If explicit offset is provided, then positions
 > start from there.  No position may preceed current offset (in other 
words,
 > leverage the rule that all elements are in order.)
 >
 > Not a big deal...just a suggestion in case you all like it.
 >
 > 
------------------------------------------------------------------------
 > Noah Mendelsohn                                    Voice: 
1-617-693-4036
 > Lotus Development Corp.                            Fax: 1-617-693-8676
 > One Rogers Street
 > Cambridge, MA 02142
 > 
------------------------------------------------------------------------
 >
 >
 >

Received on Wednesday, 31 October 2001 15:52:57 UTC