[
Lists Home |
Date Index |
Thread Index
]
I agree. It seems, however, that we do
have to reckon with the context for sending and
receiving a document, and
orchestration/choreography systems deal with that insofar as
it can be done a priori. A
dynamic builder for such systems might be a follow-on to
applied pragmatics. I've long held the
web is just big frikkin' mail system, but I've
never denied that sending love letters to
major celebrities is probabilisitcally a pointless
speech act. ("You MUST love me."
"Nope. You are Musterbating again.") The
act of negotiating in a context can be
formalized and supported by tools, but there
is no obligation to consent to
negotiation. (Again, see the papers of Aldo de Moor
and where the heck is Joshua
Allen?)
If permathreads seem fruitless, note that
the reemergence of one means there is
still contention even if it is
permanent. On the other hand, we are making it through
these by dint of practice much faster (days
instead of weeks, and sometimes hours)
so at a metalevel, learning is occurring and
in my opinion, our terminology and
theoretical grasp is improving.
Pragmatics is a model we can apply.
As noted in the paper cited, theoretical
models that satisfactorily explain precise
mathematical formulations years after the
math was created are not aberrant; they
are historically common (see Michelson,
Morley and Einstein, or Kepler and Newton).
len
> Subject: RE: [xml-dev] Did Documents Win? No. Objects Just Couldn't
GetTheir Act Together.
> Date: Fri, 24 Feb 2006 08:41:43 -0600
>
From: len.bullard@intergraph.com
>
> I think it comes down to explaining that network definitions
>
(verbs for schlepping stuff) are never *meaningful*. It's like
> asking
your mailman to do your taxes instead of moving the
> form to the IRS
and bringing the payment back. (We may have some
> fun later merging
this thread into Pragmatics (not what
> Box is talking about but the
subfield of linguistics)).
>
I think that answers the original question - to the extent that we are
converging on something like a Postal Service architecture, documents won --
the contents of an envelope are what matter (are meaningful), not the delivery
mechanism. That means that neither side in the original REST vs SOAP
debate "won":
- Protocol neutrality is important - you don't really care whether the
envelope came by plane or dogsled, UPS or USPS; it's just a cost vs
quality of service issue. (score one for the "HTTP is just a transport
layer" crowd)
- A globally understood *visible* naming / addressing system matters
a lot (score one for REST)
- A document can be just a document, a resource representation, a service
request, a service fufillment, or whatever. (Take one away from the
original SOAP/RPC style, score one for either REST or SOAP/document-literal
style)
- POST is the only externally visible "verb" that matters (score one for
SOAP 1.1, I suppose)
- Delivery confirmation (aka reliable messaging) has to be layered on top
of the delivery mechanism, and it's a pain if you have to check for yourself
whether your envelope made it thru all the intermediate hops (score one
for WS-*)
- Security comes via a complex combination of everything - transport-level
physical security, secure packaging, encrypted content (not sure what that
means for our little permathread, but at least the WS-* people realize that
security is a hard problem)
- Semantics is a private understanding between the sender and receiver;
it's none of the postal service's business what the contents of the envelope
"mean", if anything. (One more reason to be skeptical of the more grandiose
SemWeb visions, I suppose)
- At the end of the chain, it's the receiver's job to figure out what to do
with the document, whatever the sender's intent, whatever the document's
conformance to some format specification, whatever "object" was serialized to
produce it, etc. (Score one for Walter Perry).
| 