The Finetuning Manifesto.

I Find Karma (
Wed, 4 Feb 1998 16:41:54 -0800

Great picture of Bill Gates with cream pie on his face at

But this post isn't about that, it's about finetuning's manifesto.
Last night I dismissed it a little too readily. On reread, I think
that yes, Rohit, she's got a cool vision in mind.

I'll learn to be less cynical in the future. It helps that she praises
our "XML Recommended Reading" pages that I worked so hardly on. The

> Fighting for truth ...justice....and a universal syntax...
> Reinventing the Web using existing Protocols
>'s recently-revised mini-manifesto
> Okay, on to the next phase of clarification. Everybody's read up on
> all the specs, right? :-)
> Any large well-designed system for doing anything can also be viewed
> as the sum of its parts. Although ultimately mechanical in origin,
> the Internet as it exists today is the epitome of the ultimate ongoing
> experiment in scalable, dynamic, constantly refined living
> heterogeneous industrial-information system. That huge infrastructure
> when people talk about "the Web" "out there" is really an interwoven
> componentized systems of data, documents, and semantic information
> (processing instructions, scripting, interface definition mapping,
> PERL expressions, stuff like that...).
> All these elements can be combined and "triggered from" one anothers'
> actions and query responses. "If and only if value="9:00am" send
> notification", and the repetitive stuff can be automated and there are
> actually more and more standards-based custom-code generating tools
> all the time.
> How about an RDF-enabled remote search agent (your data, their
> CPU's). How much could an organization charge for the usage of that
> system's services?
> For what tasks would it be better to use a mobile data filter-bot?
> When is the "agent in the plain brown wrapper" a better buy for the
> buck, as long as its paperwork was in order.
> Web native distributed computing components will be used by
> everything, for everything. We will be making lists and checking them
> twice and then using them over and over again, where applicable. We
> will be, using digitally-signed, extensibly-designed, well-described,
> meta-data fortified, proactive, event-driven, self-describing and
> self-validifying, yet ultimately confined to the world that exists
> from within a set of given parameters whose value ranges are within
> your direct control.
> These event-driven or value-range-driven parameters and ID tags are
> designed to work heterogenously across platforms and can be encrypted,
> appropriately modified, or temporarily anoymously assigned on demand
> if necessary.
> Okay. Now how about an equally-extensible, machine-readable version so
> my "smart appliances" can interact with each other? Then my car alarm
> can radio my microwave with a recommended change of entree for dinner
> due to newly expected arrival time (increast DEFAULT "car retrieval
> time" average value per my current urban tracking coordinates and any
> other available resources). So the car alarm works it out with City
> Hall's records and the Johnson Controls utility butler for a late for
> dinner, and confirms the execution of its instructions and/or verifies
> any impending expected response values (say from a form submission).
> Oh yeah, and all this comes with even the very lowest of low-level
> processing instruction-sets. The ones that are phisically-embedded,
> hardcoded, into the chips that run your 266 MHz Pentium, reallocate
> the global-resources of an international corporation, or assist your
> respective agent-fridays to collaboratively decide to issue an extra
> reminder beep to call your mom on her birthday (based on the recorded
> sequence of events believed that proved ineffective for last year's
> agent).
> How bout banking from your car over secure and encrypted sHTTP? Or how
> about running your own cellular server right out of your automobile,
> via any number of dedicated wireless systems? Okay, you can do that
> too! It's a stupid idea, but it's possible, and half of the skills
> required for your company's survival through the impending digital
> info-wonder-circus will come from the ability to at least determine
> what you don't want for your network.
> Sometimes the practical support and maintenance considerations alone
> can determine the technology of choice using a simple process of
> elimination. And sometimes you'll watch that same simple process shoot
> your silver bullet right between the eyes.
> We can potentially transform our current Internet into a truly Web
> native distributing computing platform. HTTP 1.1 is growing like a
> weed, although perhaps still in its infancy....(yeah it's over there
> in pre-school with XML and Java -- and soon it will be in its terrible
> twos. Boy time flies when the entire multinational multi-orbital
> solar-nuclear meta-information empowered, mega-coporate and
> defense-complex financed future of "earth" is as stake, huh?
> Sounds like "Flash Gordon" kinda.
> For better or for worse: it's really happening. Whatever it is, a lot
> of really important and powerful interests are very much interested in
> participating in the impending information revolutionComing soon down
> the pipes of your town!
> What technologies are involved?
> You guessed it: XML, Java, HTTP 1.1 XML-enabled Access Control and
> Permission System, HTTP 1.1....then theres a whole mess of business
> logic rules that can be automatically generated (DTDs, Schemas,
> external scripts, embedded or internal entities, labelling systems,
> automated Unique ID systems, access and file locking mechanisms,
> semantic firewalls, information whose value increases proportionally
> with the raise in the appreciated value of the information after
> further analysing its own intradata relationships. (Seems only fair if
> some systems learn faster than others :~)
> Organic Information Systems then, are a component-based way of
> perceiving a larger heterogeneous "being". One way this kind of system
> behaves like a single beng is by demonstrating what is described as
> "self-organizing" behavoir. People will still enter information and
> run applications much in the way they always have. It's the way that
> information is processed and distributed after, sending the form for
> instance. Where does that data go? What else could that data be doing?
> Who's the worst possible person that could get at that data? How can
> you make doubleplus sure that that could never happen?
> The risk scenarios vary in bloom and color (depending on the origin of
> the weather man or fortune teller), but the holy grail remains the
> same: to enable a system capable of providing true interoperability
> between past, present, and future environments.
> The Web may eventually transform itself into a huge, unified, super
> repository of references, abstracts, research, media, and all sorts of
> corny retail and service-oriented applications as well (more than the
> 255 on your TV Screen). But the Web is not a static one-way
> content-delivery system. It's a static two-way delivery system with
> dynamic capabilities that's learning fast, and starting to talk to
> itself at this point, or go looking for answers, or figure out the
> answer itself from it's available resources.
> Oh now I've up and gone into that hazy territory of abstraction
> without meaning to. I was just trying to personify the component
> "characters" of distributed computing, and they end up sounding like
> freeliving, extremist, beatnik truth seekers. Remote agents needing an
> anoymous encrypted repository,for the night (until more system
> resources become available--the check's in the mail...)
> Here's another scenario with novel appeal because it involves a
> familiar, uniquely-customizable yet universally identifiable,
> component-based friend of our distant youth.
> You guessed it. the LEGO.
> The idea is that if the web is like a LEGO set, and there are only
> many of those neat 8 sided bendy pieces, you want to make sure that
> some appropriately designed, scalable, compatible and extensible
> architure to all of the other LEGO pieces that are going to be build
> around those pieces.
> And how can we make more of those cool pieces anyway?
> But for now, or God's sake, let's at least clean up our HTML!
> Don't mind me.I'm just trying to bring the physical and logical
> structure of the World Wide Web (From your desktop to your frontdoor).
> This manifesto itself is an organic document (in more ways than one
> :-), and will be updated with frequent certainty.


Legos, Playdoh, and Crayolas should be part of EVERY household.
-- Ari Rapkin