OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   [permathread:semantics] What Markup Is For

[ Lists Home | Date Index | Thread Index ]
  • To: XML DEV <xml-dev@lists.xml.org>
  • Subject: [permathread:semantics] What Markup Is For
  • From: "W. E. Perry" <wperry@fiduciary.com>
  • Date: Thu, 10 Apr 2003 11:57:10 -0400
  • Organization: Fiduciary Automation

At the crucial moment of his argument in a piece called 'On Semantics
and Markup'
http://www.tbray.org/ongoing/When/200x/2003/04/09/SemanticMarkup Tim
Bray strikes a pose of Socratic agnosis with:  "To oversimplify, XML is
winning and ASN.1 is losing. There are a variety of reasons for this,
but one of them is that it seems to be more important to know what
something is called than what data type it is. This result is not
obvious from first principles, and has to count as something of a
surprise in the big picture."

To be candid, there is nothing surprising here if we understand what a
label is, and that it is in fact a direct consequence of first
principles that labels are the required input to processes which
elaborate semantic outcomes, but typing of any sort is not. Quite
simply, what datatype something 'is' is a corollary of that something
being manipulated as of that type by a process. Type--and not just
datatype--as opposed to labelling, inheres in process and not in the
text or data submitted to processing. Interoperability, which to the
extent that term is properly used to signify the transmission of real
semantics, or meaning, is the essence of communication. It is
accomplished by noun, or label, overloading. A label used in one way and
as of one type by my process and in another way and of another type by
yours permits us both to do useful but different things--that is,
elaborate different semantics appropriate to each of us--over the same
labelled content, text or data.

Such labelling is what markup is for. For too long in data processing we
have been accustomed to the two-phase commit model of processes sharing
an identical data structure. Being forced to share datatypes as well as
labels severely curtails the specificity and expertise of what a process
might do with given input. Though those restrictions might have been
appropriate in the homogenous network behind the enterprise firewall, we
cannot now ignore the implications of direct any-to-any node-to-node
communication across the internetwork. Type is everywhere inherent in
process, not in content, but on the internetwork the specificity,
autonomy, locality and terminality (SALT, not ACID) of processes
preclude them from sharing local type as a basis for non-local


Walter Perry


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS