This is a Wiki type which is composed of more than the pages which stand as its base - c2.com
(The following Thinking Out Loud is being revised as several Personal Information Collections via several processes (Nys Lte, Nys Artifacts and a revised Nys Wiki Server) which are presently (2013-2014) being developed, compiled, presented, celebrated, and reposited at a web site (sunflowersynergies.org ). Containing some or all of the following:
Wiki Features installed in a locally run wiki application which adds functions normally not included or made visible in simple wiki applications.
(This applies to development of this new class of wikis and is not suggested for any existing wikis.)
Page Status -- specific changes in user-selected pages including
The client must be able to utilize abilities built into the server, which will also enhance More Than Page Functions which will generate useful information.
This sounds somewhat similar to an idea I'm playing with in my head which I'm tentatively calling 'Dataspace' (for want of a better name) at my journal at natecull.livejournal.com. The idea is basically an executable Semantic Web, used as a single replacement for a publishing system and desktop: instead of dividing data into filesystems and process trees, everything becomes a single 'executable memory space' of live data cells: something like a World Wide Spreadsheet (but a bit more general than any existing spreadsheet, SQL database, or RDF implementation). It is a concept that seems like it ought to be incredibly simple, but I'm struggling to try to map it onto existing languages or programming paradigms. Of all the stuff I've read so far, Ted Nelson's 'Xanadu' vision resonates the most (particularly his ideas about applications becoming personal hyperlinked information spaces called 'applitudes'), but I can't follow him entirely. Wiki also approaches the idea pretty closely, but it needs to become editable at a far smaller granularity than the 'page' - something like a global space of interlinked personal node-based wikis, where each node can have and be metadata for other nodes, and be seamlessly updateable in realtime as data sources change, is closer to my vision. This ideally would scale both up and down from a Semantic Web type global structure to a fast, lightweight machine-level parallel-programming framework.
My main motivation for exploring massive data-parallelism is not raw speed but a sense that a) data representation is just naturally parallel and we should not introduce artificial sequentialism or it will break the simplicity, b) most things we're trying to do in programming are just not that hard and life really should be a lot more pleasant than it seems it actually is right now. GUI programming in particular seems like it hurts far, far more than the simplicity of the underlying ideas needs to, and c) what we want in an age of the Web is massive democratization and massive data remix across multiple data sources and services; that means eliminating the difference between 'personal data storage' and 'personal applications/desktops', eliminating the difference between 'writing a program' and 'installing a program', and building something more like 'personal data dashboards' that receive, mix, blend and apply data and functions drawn equally from all over a globally distributed, locally cached utility Web, and that a reasonably interested 'user/programmer' could safely create without messing up the rest of the world, much like HTML enabled the creation of the first Web.
Of course, the hard part is trying to grasp the right abstractions for the first building blocks, and that's where I'm unstuck at the moment. I know I need some concept of a 'hierarchical namespace' (or dictionary) and a fairly simple 'publish, subscribe, compute' framework and a basic language/syntax, and it should be rather counterintuitively both *hardcore* pure-functional and *not* strictly typed at all (which rules out Haskell), but getting all the bits together to make the simplest possible working model still eludes me.
(It seems like RDF *ought* to be able to do this, but RDF has a lot of Missing Fundamental Pieces - primarily an actual programming *language* with Data Parallel Semantics that can itself be expressed in RDF.)
If anyone else has any idea of just what this paradigm is, because there must surely be a name for it, I'd love to know - it would save me a lot of effort if someone else has already built it.
-- Nate Cull
You're all, without knowing it, groping towards the ideal of the Glass Bead Game
Thanks for the link... I sent a note to him. -- Mark Janssen
# See also