[
Lists Home |
Date Index |
Thread Index
]
- From: "W. E. Perry" <wperry@fiduciary.com>
- To: xml-dev@xml.org
- Date: Mon, 01 May 2000 15:04:31 -0400
[My last two postings on this topic have not yet come to me from xml-dev, though this
much later one from Eric Bohlman has arrived. I shall reply to it here and hope that
the out-of-sequence thread which may result will not make the topic impossible to
follow.--w.p.]
Eric Bohlman wrote:
> Human communication (and organizations are made up of humans) has *always*
> been based on shared context. The symbols humans use to communicate
> derive their meaning from this shared context, not from any intrinsic
> properties of the symbols themselves. Human language is negotiated, not
> received. Dictionaries *describe* what meanings people *create* for
> words, not *prescribe* usage based on intrinsic properties of
> words. Attempts to create artificial human languages have failed because
> they don't take this into account; if there's a language barrier between A
> and B, it's only going to be solved by *joint* action between A and B
> (each trying to learn the other's language, or the two creating a pidgin),
> not by C acting at a distance.
This is perfectly on-target IMHO. The present opportunity to break free of the pre-
and identically-defined data structures of the two-phase commit model depends on all
three of these characteristics: The shared context will be created by the exchange of
messages between previously anonymous autonomous nodes. The nature of the transaction
will be negotiated in the intrinsic terms of the transaction itself. The history of
transactions which any two nodes may build between themselves will amount to a
descriptive repository of the forms into which they have successfully instantiated
each other's data in order to consummate particular transactions.
None of this is pie-in-the-sky, or even particularly difficult. All of it is greatly
facilitated by inherent characteristics of XML and ot the Internet topology.
Implementing it requires, first, the will to break away from the current two-phase
commit understanding of transaction processing and transactional data structures.
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/
***************************************************************************
|
