memerising

Tinderbox is now out for MacOS X! Yay, no more reason to run Classic! Victor Lombardi has a nice article on how he set up Tinderbox to do his weblogging. I'll have to see how to combine Radio Userland and Tinderbox. Motivation: Tinderbox provides personal inf-formation management, with multiple 2-D palettes and views, through which I can project to public and shared spaces (both those I receive and those I contribute to). Ie, how to layer my personal ontologies (vocabularies of significance) on larger shared ontologies (public memes).

RESTful coupling as the nature of the noosphere.... REST (REpresentation State Transfer) is a URI-centric architectural pattern for the web, featuring information nodes addressable through URLs, and request semantics fully expressible through HTTP GETs. This is contrasted with an API-centric approach (SOAP) which uses the web purely for transport, with its data unpacked by some language or application specific API behind the URL. The difference may seem subtle, but it's significant: REST allows the language of the web to grow. Its semantics are within the ontology of the web, rather than within the ontology of some service behind the web. What this means is that the REST pattern can allow rich growth of meaningful links among web nodes: it can enrich the language of the web. You offer verbs with visible structure, somewhat like RNA, which guide the building of higher-level structures. This is synergistic with Loose coupling as the way to go with growing the semantic web or the net of web services. You have nodes which expose services, or act as clients for services. Out of these you weave living info-ecologies. Not single monolithic applications, but information-husbandry patterns that you and your software can grow. This all needs common standards and use-patterns. The key pieces seem to be:

addressable nodes (such as purple numbers)
metadata on the nodes (both embedded and layered on)
metadata on the links
persistence of nodes (you need to evaluate persistence-trustiness, and cache/archive if necessary)
cascading ontologies (You always want to add your own metadata to that offered by some group or topic specific metadata management service).

Blogging with backlinks is an example of enriching through RESTful coupling: add your metadata on an inf (an information item, in RESTful space, that you're commenting on in your blog), which becomes part of that inf through its backlink to your blog. It works both ways: the weblog item points to and enriches an inf, and the infs pointing to the weblog item enrich it. Cory Doctorow has a nice discussion of this in My Blog, My Outboard Brain .

An evolving architecture of participation: Tim O'Reilly said it really well in a talk at Apple's World Wide Developer's Conference:

In a talk titled, Watching the Alpha Geeks, O'Reilly stated that disruptive innovation is again changing the computing landscape, as innovators and hackers are using the Web, wireless communication and built-in services to push aside the old paradigms. ... O'Reilly painted a picture where there is less focus on ownership and more on letting the user determine an evolving "architecture of participation."

Persistent URI nodes become magnets for gathering together like-minded souls with appropriate workspaces and services. Contexts (symbiotically maintained through links, relationships, metadata, topic maps) enrich nodes with meaning. A great example are the referrer tails (refurry tails) being grown by some nodes: cf Jon Udell's Backlinks in Blogspace

The culture of blogspace is evolving in near-realtime. Last week, a new mutation brought backlinks into a more prominent role. At Disenchanted, inbound links were automatically reflected outward. Each article grew a tail of backlinks that pointed to pages referring back to it. Suddenly a new kind of feedback loop was created. With a twist of the lens, conversations that had been diffuse and indirect came sharply into focus. Almost immediately the meme replicated.

Jack Park posted a pointer to Anticipation on the Bootstrap Alliance's Open Hypertext list. Which is right along the lines of call-by-future, as noted in my response (this link directly to the relevant paragraph courtesy of purple numbers). call-by-future is an argument-evaluation and existentialization pattern offered by some programming environments, related to lazy evaluation (complementing call-by-value, call-by-reference, etc). Ie, we can proceed in confidence that we'll get the values we need when we really need them. This makes sense in highly distributed and asynchronous systems. I argue that this is related to the notion of intentionality in the context of the binding problem, how previously disparate perceptions and actions are unified into identifiables. Ideally, we provide the call-by-future for our symbiotic software, co-emerging as the living web of the Noosphere. Homage to Teilhard de Chardin!

Presenceworks " integrates the presence from instant messaging" (including all the major services - ICQ, AIM, MSN...) to your applications. Their server "streams the real-time online/offline status of your contacts directly into your systems." Leave the chatting/IMing to peoples' favorite clients: just make their presence info available to each other in handy, info-nurturing group environments.

David Reed on GFNs (Group Forming Networks): " Networks that support the construction of communicating groups create value that scales exponentially with network size, i.e. much more rapidly than Metcalfe's square law. I will call such networks Group-Forming Networks, or GFNs." Just don't think you can exponentially monetize this :-)

This Doonesbury cartoon reminds me of the virtuality of reality... We're already co-creating and making it up together - and that's also what the web is about. Or, as Aretha Franklin said, Who's Zoomin' Who?... :-)

I've been watching how my 9-year-old son is making use of the Lego Mindstorms community and associated sites: he's self-educating himself, making use of both books and online resources. The lego robotics forums let him see what builders all over the world are constructing, complete with digital photos of construction details. He refers to these in his own building projects, always with variations due to different parts, etc. He can go to the Lego Store and find products (using a cool cross-hairs type interface), zoom in to the actual package (front and back), zoom inside to see how the parts are packaged, and then see a structured listing of the parts, so he can understand what he'll get and how that fits into his current inventory of gears, motors, sensors, pneumatics, hydraulics, etc. Pretty inspiring! He started out with the USB Team Challenge Pack, which includes one programmable RCX brick (3 outputs (eg., for motors) and 3 inputs (for sensors)), and the Robolab language (a visual [example] programming language co-developed by Tufts University and National Instruments and based on LabVIEW (which was used to control the Mars rover)). You beam the program to the brick by infrared. This is education/collaboration of the future, in practice, now.

Well, Jon Udell's article on Instant outlining deserves attention. Instant Outlining: like Instant Messaging, in that nodes (messages) can be shared to those subscribing to them, like e-mail, in that messages are shared between people, and like blogging, in that you get a narrative of ideas. But more, in the potential for semantic links and structure... Let's intertwingle!

Here's Tim Berners-Lee with a pretty good high-level view of the Semantic Web: interestingly enough describes it using a scenario similar to that in Apple's 1987 Knowledge Navigator concept ... basically you have your personal info husbandry agents which are like RDF-fixing bacteria nurturing URI nodes - my image! He says, "With RDF, you write a few rules, and sharing the data is a breeze. It works as a hub, connecting your applications as spokes." Ie, in a URI-centric world, you want to describe the relationships between the nodes (this arc connecting those two nodes is an authored-by link). Sense-making... The nodes and the arcs are rich with metadata describing relationships. Since this web is symbiotic with us humans it must be human readable (so our spontaneities can move it), and since it fosters autonomous agent activities it must also be machine readable.

Tinderbox looks really exciting - Basically a 2D graphic semantic mapper, so you can place your nodes (documents, links, RSS feeds) and their relationships, and also publish out to weblogs. Classic Mac only at the moment, but OSX (and Windows) coming soon. I need something like this to arrange my stuff in a space with muscular (where to reach out for that inf) and semantic memory. I used to run Web Squirrel, one of their previous products. Sophisticated, but wasn't handy enough for me to keep using it. Similarly for The Brain... Another one I'm watching is Six Degrees (due to arrive in July for MacOS X and Windows): "rather than replace existing tools, it looks at the relationship between files, messages and people and creates connections. It presents the user with all the information about a project, regardless of where the project files or messages are stored, and without imposing a different way of working on the user. " All these products have potential to start making sense of our scattered infs (information items), including how to sediment linear streams (usually temporal, but more generally storage-ordered rather than intention-oriented) into more comprehensible and articulated personal geographies. It sounds like that with Tinderbox you could have a personal info organizer (a non-linear geography) with RSS streams coming in and weblogs going out. Tinderbox is XML based, so it could be open (to the real world of inf-formation) and interoperable enough to be a real winner. I've so far managed to stick to OSX only, but giving Tinderbox a try might be reason enough to run Classic.