[
Lists Home |
Date Index |
Thread Index
]
1/13/2002 7:43:43 AM, Sean McGrath
<sean.mcgrath@propylon.com> wrote:
>
> Nature figured this out long before we
> sentients did. Extreme programming/well-formed XML/
> procedural scripting is a good toolset to start
> with to mimic natures ability to grow complex order
> out of large assemblies of simple interactions. In nature,
> powerful functionality emerges from the bottom up.
> The queen in the ant hill is not a monarch. There is no
> "top down management" and the functionality
> did not emerge from a top down design.
Sean may be reading the same stuff that I am lately! Maybe
"Emergence: The Connected Lives of Ants, Brains, Cities, and
Software" by Steven Johnson .... Also, I'm not sure what
more I can say about this (being tangentially involved), but
people might want to look at
http://www.amazon.com/exec/obidos/ASIN/0201776413/qid=
1010934914/sr=1-2/ref=sr_1_10_3/002-3841261-8649657 when it
is published in a couple of months. Do you have any reading
lists to suggest along this line, Sean?
Anyway, I'm trying to sort out in my mind a "thesis" that
goes something like this:
XML was conceived as "SGML for the Web," combining aspects of
both its SGML heritage and its Web heritage. SGML puts the
Document Type Definition at the center of an application;
design begins with a document analysis, proceeds through a
detailed DTD design, and ends up with application components
that are highly designed and coordinated by the "authority"
of the DTD. In other words, the DTD is sortof the "queen
ant" (in the old-fashioned sense) whose authority keeps it
all together. XML's formal development within the W3C has
tended in this vein, though of course a typed XML schema and
PSVI is the "queen" of a state of the art XML application.
This works fine in "behind the firewall" environments (or
really well organized industries) where the complexity can be
managed by authority. This tends to not work when there
isn't an authoritative schema (or many contending schema, or
schema in rapid evolution), and this "balkanization" or
"tower of babel" is perceived by many as a crisis for XML.
For a recent example, http://www.eweek.com/article/0,3658,s%
253D1884%2526a%253D20656,00.asp "Now, more than three years
after XML's introduction, IT shops implementing industry-
specific variants find themselves looking at multiyear,
multimillion-dollar projects that leave two fundamental
obstacles unchallenged: how to shift partners from trading
through traditional means to trading with XML and how to
interoperate with other industries. "
Of course, one solution is for customers to sell their souls
to a dominant vendor with a useable schema, or to lash the
standards groups until they come up with usable vendor
neutral schemata, but think of the other half of XML's legacy
-- the Web -- and what made it succeed without a dominant
vendor or centralized authority. That's where, as Sean
noted, the modern conception of the ant queen as simply the
mother of the colony becomes an interesting metaphor:
Anthills exhibit effective "emergent" behavior by individuals
operating very simply, the queen does nothing but lay eggs
and nobody coordinates (except Darwin, of course). Likewise
with the web, no committee laid out requirements for the Web
and reviewed designs to achieve it. Instead, its features are
the emergent properties that appeared when TCP/IP and HTTP
produced a reasonably reliable universal network and the
simple but powerful HTML (and later XML) markup languages
became almost universally supported in browsers and authoring
tools.
I'm not at all sure what this vision of an alternative,
loosely coupled approach to using XML to build complex
networks of Web applications/services implies for us, other
than "stop worrying about balkanization and seeking
uniformity," "keep the components simple" and "try lots of
things, keep the successes, and kill off the failures."
Steven Johnson's book discusses some ways to use the "ants
following pheromone trails" metaphor to improve the Web,
especially www.alexa.com, which provides a feedback loop that
HTTP/HTML doesn't have natively. I'd say that Google has done
a lot along these lines as well, as have weblogs. To the best
of my knowledge (and it is pretty much gossip and hearsay),
the Wall Street types have thrown a lot of money at the
complexity theorists in Santa Fe and gotten little return on
their investment, so I'm not at all suggesting that we
develop Strange Attractor Markup Language or takes ants and
swarms and boids too literally as models for web services.
But I do think that there is SOMETHING here, and I'm very
glad to see Sean thinking the same way ...and would be very
glad to hear from both people who think they can effectively
debunk this and from those who think they can add flesh to
its bones.
|
