xml-dev - Re: SOAP, plague, love

["W. E. Perry" <wperry@fiduciary.com>]

Eldar Musayev wrote:

> If you mean that XML is going to be overcomplicated, it's true.
> If you mean that SOAP puts semantics on XML, that's false.

>From http://www.w3.org/TR/SOAP, dated today:
SOAP does not itself define any application semantics such as a programming
model or implementation specific semantics; rather it defines a simple
mechanism for expressing application semantics by providing a modular packaging
model and encoding mechanisms for encoding data within modules.

The composer of a document, message, or RPC request *defines* the semantics (or
works with received semasiology). The (self-proclaimed) purpose of SOAP is
'expressing application semantics'. Do those words not denote that SOAP conveys
specific semantics explicitly intended to be understood as processing
instructions ('applications')? In the absence of such semantics, XML is merely
syntax; semantic baggage has been put upon it by SOAP.

>  > In my unorthodox view, the specification of XML is a
>  > specification of syntax
>  > alone. The processing of that syntax to derive some
>  > semantics is my (and every
>  > node's) purely local and idiosyncratic affair.
>
> So? And what is so unorthodox here? Of course it is.

It is apparently outside your orthodoxy, and certainly outside that of SOAP,
that I expect in every case to derive my own semantics, without reference to
the intent--expressed in SOAP, or otherwise--of the composer or sender of the
message. I might, for example, treat your message itself as no more than an
event notification. Ignoring not only its semantics, but its data content and
even most of its syntax, I might use receipt of your message as the trigger to
go elsewhere and fetch the data I actually need. I build systems just like that
all the time. An example:  A securities trade is executed and an execution
notification produced for interested parties. I have registered myself as an
interested party to that trade because I must execute an offsetting follow-on
transaction. I don't care what the content of that execution notification is,
and I certainly don't care how the sender expects that his semantics will
signal me to process it. I care only that is is completed:  I will take that as
the signal to perform some process that the sender of that message knows
nothing of, and has neither the semantic vocabulary nor my attention for
instructing me how to process his message. This is quite typical of how my real
world works. It may well be unorthodox. Transactional events produce outcomes
which others, often unknown to the original transacting parties, may then act
on. They act out of their own understanding of those original outcomes and from
their own expertise in processing them further for their own purposes. The
syntax of XML (i.e., XML as defined) is beautifully suited to expressing the
informational content of what has occurred--the outcome, that is, which now
exists. The interoperability of the message expressing that outcome lies mainly
in the clarity with which it can be understood, and processed, by further
autonomous nodes for their own purposes.

> It's just when you start to implement and especially
> integrate huge B2B systems, you have to share those idiosyncrasies
> with more and more of your business partners, so they can understand you.

No, I don't. The idiosyncrasies are in what I *process* privately, based on
messages openly exchanged.

Respectfully,

Walter Perry



***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************