[
Lists Home |
Date Index |
Thread Index
]
On Tue, 25 Feb 2003 13:12:12 -0600, Bullard, Claude L (Len)
<clbullar@ingr.com> wrote:
> So I ask again:
So I answer again :-) [This being a permathread and all]
>
> 1. Why does SOAP need a subset sanctioned at a higher level than its own
> specification?
SOAP doesn't need a subset. Users of XML tools building documents they
expect to send via SOAP, or building SOAP envelopes for their XML
documents/data "by hand" might need tools that enforce the SOAP subset so
that some downstream SOAP processor doesn't reject them.
>
> 2. Would the XML-SW serve that requirement whatever that requirement is?
Not as both are currently specified, AFAIK. Net the DOCTYPE and PIs, I
would think they are within striking distance of one another. Should they
try to work out the differences? I dunno ... the XMLP WG is almost
certainly not interested in re-opening the issue at this late date, yet I
couldn't make a plausible argument why XML SW would deprecate DOCTYPE
declarations or PIs.
Had XML SW existed a year or two ago, I suspect that the SOAP people would
have gone along with it; I can't remember how fervent the opposition to PIs
inside the BODY is. From my point of view, PIs would be unmitigated evil
in web services if they were used in SOAP headers as they are in documents,
i.e. to denote out-of-band private contracts between the producer and some
specific processor. That's merely tedious in the document world, but could
be fatal when interoperability is the sine qua non. [Just my personal
opinion, I can't remember the XMLP discussions very well and it's never
come up in WSA].
But as tricks such as namespace declarations with a private URN illustrate,
you can reference an out-of-band private contract in all sorts of ways
besides PIs, so I'm not sure the fight is worth having. I think XML-SW
would be a Good Thing for almost everyone including the Web services
people, although it might cause a bit of discomfort to align with it.
|
