Re: proposal for eliminating message

Sanjiva,

your proposal has several parts to it (if I haven't identified them all,
do correct me, please):

     1. make part name optional
     2. add cardinality indicator
     3. remove the <message> syntactic construct (together with a symbol
        space)

I trust these points can and should be considered separately.

I don't see the need to make part names optional, but then it's not
clear to me at the moment why they were mandatory in the first place.

Adding cardinality indication is wanted by many, but then they (or
others) cried that reinventing XML Schema is bad.

Finally, I don't see any point in removing the syntactic construct. I
believe the resulting syntax is more confusing *and* we lose the ability
to reuse message definitions in different operations. I don't like this
part.

So my votes on the three points in order would be: dunno, yes, no.

Now because you haven't removed the construct, you've just moved it
elsewhere in the syntax, I don't think you're going to satisfy the
opponents to <message>.

Until I'm persuaded mixing existing type systems (e.g. DTD or MIME in
XML Schema) doesn't have ugly consequences, I'd keep our simplistic flat
message container specification.

Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet Corporation
                   http://www.systinet.com/






On Sat, 2003-01-18 at 08:30, Sanjiva Weerawarana wrote:
> Attached is an attempt at a compromise proposal for removing the
> <message> construct.
> 
> Sanjiva.
>

Received on Tuesday, 21 January 2003 13:38:57 UTC