~ilmu/tala.saman

This is the reference implementation of datalisp I suppose

b2a5821 make it a bit clearer how we intend to tackle incentives of coordination

~ilmu pushed to ~ilmu/intro git

25 days ago

f1c00b7 fixed unbalanced parenthesis in output to collect

~ilmu pushed to ~ilmu/cxp git

26 days ago
Just as the web was made to exchange hypertext, datalisp is made to exchange behaviours.

Our goal is to index all useful behaviours of computers (forming an open cover) using a
decentralized namesystem built on canonical S-expressions (interacted with via menus).

Basically a package manager thing.

Ok so, "general purpose metaprogramming with gradual types" is one way to see it
the mvp is a "gradually typed version control system" although there are other 
perspectives that are insightful, such as focus on authenticated datastructures.

We want to relate properties to objects using various means, most notably probabilistic
estimates of trust, but programs as predicates is also a heavyweight. These concept 
lattices are one of the interesting interpretations of hypergraphs w.r.t. computation, 
others are described in papers. The properties will then be related logically with 
datalog-ish expressivity (the general architecture looks a lot like datafun). 

Fundamental motivation of the project is to combat pollution; the tragedy of the commons.

We are interested in building society on a stable foundation and with computers 
becoming an increasingly fundamental part of the coordination of human behaviour 
we worry about the incentives being created. By building an index in the space of
coordination-free programs (the datalog-ish expressivity) we can then use proven
methods to coordinate (at the edge), thereby straddling the boundaries of domains.

The datalisp project builds on some very simple ideas: 
- the duality of code and data (universal turing machine proves this)
- canonical representatives for equality comparisons (think torrent seeding)
- free software as a foundation for a democratic society
- the peer to peer model as a more accurate description of reality
- therefore; decentralized coordination of society
- persuasion as means to updating software (coercion / centralization is systemic risk)
- trust as a channel for effective communication (information you don't trust is noise)
- probabilistic assessment of trust as the foundation of economics
- sufficiency of economics to coordinate name system
- sufficiency of name system to coordinate society
- conflict resolution as a "join" in name system lattice (actually is the meet - trade)
- reified conflicts as a form of conflict resolution (the join - market)
- currency exchange rates as measurement of relative legitimacy

Feel free to contact me with questions or join us on matrix: 
https://matrix.to/#/#datalisp:matrix.org