d∈f.coδes
Say what you mean
def.codes is an independent research system for exploring the space of graphical, knowledge-based modeling using web standards.
You can reach me at g@def.codes
Major concepts
- RDF the ultimate bytecode
- a ubiquitous userland knowledge base
- “propositional programming”
- open-world interpretation and composition of knowledge graphs
- vocabulary design for monotonic semantics
- atomic, non-monotonic reification of world descriptions
- well-defined host ↔ knowledge base mappings
- logical induction of userland representations
- language service integration with live systems
Fields
- symbolic programming
- logic programming
- description logics
- rule-based systems
- term rewriting
- exploratory programming
Things in a world
def.codes adopts a graphical paradigm where computational constructs are regarded as “things in a world.” This might sound like a low bar, until you consider that the so-called “things” that software practitioners currently work with lack most of the essential properties that we expect things to have in a world. As humans, we expect things in our world to have certain basic properties before we can consider them “real things.” A real thing should be:
- discoverable
- you can find it
- tangible
- you can witness it
- explorable
- you can approach it in different ways
- portable
- you can take it with you, if small
- describable
- you can say something about it
- indicable
- you can point to it
- addressable
- you can talk to it
- operable
- you can try things on it
- interoperable
- you can introduce it to new things
- trackable
- you can ask where it came from
- recomposable
- you can take it apart (if assembled) and use the parts
Most “things” in computer systems lack these qualities. In order for them to have psychological reality, humans must substantially fill in the gaps using logic and imagination.
The thesis of def.codes is that these qualities all require a knowledge basis. def.codes is thus a system that bottoms out on knowledge representation.
System
The system requires only a JavaScript interpreter. The core package
provides an API, exposed as a module export but used idiomatically as
a global singleton (def). The API provides an interface
to an RDF 1.1 dataset.