My friend Aharon looks at SdiDesk and still finds it doesn't go far enough ...
He says : if you can write a text and see it represented as a diagram, a flower, a group of agents or whatever and be able to interact with it like that.
Text as a diagram
He wants a closer relation between diagrams and texts. Now in one sense, I know he's surprised that the diagrams in SdiDesk aren't showing the ordinary links between pages (as in TouchGraphWikiBrowser). That's partly deliberate : I want diagrams to show selective "views" on the relationships between pages. They're basically spatial indexes.
But it's partly rationalization because re-inventing TouchGraph would be a lot of work.
So how to get better integration? At some point automatically generating a diagram should be an option. But it shouldn't be the only way of treating diagrams.
There is also another issue, if you draw links between nodes in diagrams, should they be represented in pages? Maybe automatically added to a "see also" list.
Text as a group of agents
Because of the context of our conversation, I know he was refering to my previous incarnation doing ArtificialLife simulations of populations of evolving agents. How can a text be a population of interacting agents? Well, clearly some texts are programs. But in a sense, a wiki is a population of nodes which already are "interacting" in the sense of having links between them. It has high-level, emergent structure. You can do population analyses (see AnalyzingThoughtStorms). You can look at it's evolution over time (HistoryFlow). You can think about it ecologically. Niches, competition, refactoring, adaption.
But how do you interact with it as a population? Well, you certainly interact with a community of contributors. There's politics. And that's one way of treating text as a group of agents. There's WikiArchy - some kind of vote-edit relation between wikis.
And if you're thinking about virtual agents. There's always the possibility of them wandering round a wiki and processing it in some way. (AutomaticWiki) Can you send bot armies to fight WikiPoweredArgument?
Or evolve TuringTest passing contributors. Imagine bots posting what they find on the web, into a wiki, then killing off those which are accused by the human population of being mere bots. Evolving smarter, more relevant, more lingiustically capable bots?
Text as a flower
But text as a flower? That's a tricky one.
There are obvious GardenImagologies for wiki. The continuous tending, pruning, weeding, planting. There's organic growth.
And maybe you should be able to "pick" the blooms. The finest ideas for dropping into a vase (or weblog) to briefly decorate some other space before it withers and dies. (Remember IdeasAreCheap)
And there's also on of my current UnfinishedProjects (a work, so lacksadasically, in progress, that it's never even been announced) provisionally entitled FlowerBrush which could be said to be about ContestedSpaces and so linked to CompetitiveArchitecture.
But what really would be "text as a flower"?
I think it's not so much a flower as a plant... Then one just use the process, or the quality of plants... (e.g. root links feed in, atmospheric links feed in, etc..)
Look, I just had a flash thing regarding your idea of looking at thoughts and thinking in
percpectives. These are like different cultures and mythologies, aren't they? Like viewing
through dates and dating and calendaring (kind of Maya style..?), or having gods for human
processes like god of thinking, god of feeling, etc.. (kind of nordic style...?) Or like life
dream, or through animal behaviour, or through deduction, or through chancing, getting deeply
stoned, etc... So one may have a chance option, dedconstructmytext (sure, will require some
context algo.), bio-rythmical presentation of your texts, assigining thoughts for the gods, etc..
Sure, one can have too many options, but if users can twich things on as they want... Also, and
this I think could again be very powerful, is that such percpectives could be used for links
themselves... Basically, you know how you say that you tend to move from one thing to another,
and do not seem to be in sync with what you supposed to be doing? Isn't it that you use a
different percpective to kind of counter another? Maybe with this kind of program, you can
combine the colours of your rainbow into a certain form?
Sorry, am running with it a bit..