Dear Graham,
Well, I get where you're planning to go but i) the rationale is somewhat trivially showy and you seem to have both ii) seriously overestimated the capabilities of the current technology and iii) seriously underestimated the scope and scale of the challenge (to a degree that forces me to question the depth of your understanding of the domain).
you didnt give much detail here, but on the rest of your comment, so i can't answer to this part.
i) You need to provide more support for breathtakingly-wide statements such as This class is isomorphic to the class of intuitionistic proofs. (Can you prove it?)
those are well known results. cf
http://en.wikipedia.org/wiki/Intuitionistic_type_theory for example
ii) Let's take
Déductions as an example, what is
actually available is: ... rules in N3 language to generate simple Create-Update applications for Java Swing platform from OWL, RDFS, or UML ... That's parsecs away from the capability that you're assuming will be available to underpin your modelling.
by any means, i cannot see how automatic GUI building is more than a piece of cake. I've done such myself zillion times, for example, i already built automatic GUI with QT for my first sketches of zennet.
iii) The complexity of the modelling task militates against anything less powerful than OWL Full and that effectively renders the task impracticable with contemporary reasoners. And - you'll find out for yourself soon enough when you get down the nitty-gritty - all this nonsense about bolting ontologies together is just that.
I do take into account that some logic algos here aren't trivial at all. Some will use SMT solvers, some are only human solvable, and some are not solvable or not known to be solvable. Still, many operations like consistency check can be done efficiently.
BTW, you should probably include in the references Jezza Carroll and Chris Bizer's original paper:
Modelling Context using Named Graphs. (And you should note that adopting this tactic for modelling context then obliges you to introduce and maintain an explicit temporal representation (of now) which is going to pose its own profound problems for the reasoning).
tx, nice paper, maybe will make a reference. wrt time ordering, ofc Satoshi ideas have to take place.
I will say that your approach is at least broadly coherent, as opposed to merely spouting risible gobbledygook like the other altcoins that purport to wave an AI flag.
$2m won't cover it though, $20m just might.
Let's keep the discussion going and see where it takes us. I cannot answer such a vague claim, ofc. You sound like you know what you're talking about, and I'm very happy to learn.
