extensibility, open exchange

["Simon St.Laurent" <simonstl@simonstl.com>]

A large portion of the XML community sells XML softly, as a set of tools
for defining (often programatically enforceable) standards representing
the terms of agreement for communication between participants.  

XML also has a core of people who take "extensible" more seriously, and
regard XML as a syntactic foundation for communications of a rather more
anarchic sort, in which participants can use whatever labels and
structures they want to send information to other participants who read
those labels and structures however they want.

The former group commonly regards the latter group as far-out dreamers,
while the latter group often regards the former group as lacking in
imagination.  (Both of those descriptions are substantially toned down.)

For the far-out dreamers (and those willing to listen), I've got a brief
quote that explores the possibilities created by the model of open
exchanges.  The quote is specifically about the nature of scientific
debate, but it isn't difficult to stretch "collectively deciding an
issue" to other forms of communication, even markup processed in deeply
'rational' computing environments.

"There are therefore at least two different ways of collectively
deciding an issue which I shall call a guided exchange and an open
exchange respectively.

In the first case some or all participants adopt a well-specified
tradition and accept only those responses that correspond to its
standards.  If one party has not yet become a participant of the chosen
tradition he will be badgered, persuaded, 'educated' until he does - and
then the exchange begins.  Education is separated from decisive debates,
it occurs at an early stage, and guarantees that the the grown-ups will
behave properly....

An open exchange, on the other hand, is guided by a pragmatic
philosophy.  The tradition adopted by the parties is unspecified in the
beginning and developers as the exchange proceeds.  The participants get
immersed into each other's ways of thinking, feeling, perceiving to such
an extent that their ideas, perceptions, world-views may be entirely
changed - they become different people participating in a new and
different tradition.  An open exchange respects the partner whether he
is an individual or an entire culture, while a rational exchange
promises respect only within the framework of rational debate." [1]

Open exchanges have historically been expensive things in computing
terms, as the huge sums of money thrown to relatively weak
implementations of natural language processing have demonstrated.  For
some of us, however, XML offers an intriguing middle ground, bringing
structures which lower the cost of processing while offering
capabilities which are far more open than constrained vocabularies can
ever be.

Thanks to Michel Biezunski for mentioning Feyerabend in a conversation a
few weeks ago - I'd always known it was there, but it's been fun reading
ever since. 

[1] - Feyerabend, Paul. _Against Method_. (Verso: New York, 1993.)
p.227-228.

Simon St.Laurent
http://simonstl.com