3 ms·

Very interesting! I especially like the expressed interest in logic-based approaches for reasoning about graphs. Quoting from the article: "Logic-based and dec

by triska 5y ago

Very interesting! I especially like the expressed interest in logic-based approaches for reasoning about graphs. Quoting from the article: "Logic-based and declarative formalisms. Logic provides a unifying formalism for expressing queries, optimizations, integrity constraints, and integration rules. Starting from Codd's seminal insight relating logical formulae to relational queries, many first order (FO) logic fragments have been used to formally define query languages with desirable properties such as decidable evaluation. Graph query languages are essentially a syntactic variant of FO augmented with recursive capabilities. [...] The influence of logic is pivotal not only to database languages, but also as a foundation for combining logical reasoning with statistical learning in AI. Logical reasoning derives categorical notions about a piece of data by logical deduction." This interest could be good news for vendors of logic programming languages such as Prolog and its syntactic subset Datalog. These languages enable very convenient reasoning about graphs and allow both storing and querying the database with the same simple language elements that can themselves also be analyzed and reasoned about with the same formalism. Imagine how convenient it would be to have a large graph database represented as a set of Prolog clauses that can be queried, run and analyzed with an efficient Prolog system! The data itself, queries, and any desired transformation and augmentation could be uniformly expressed in this way.

sfvisser 5y ago
This has been the semantic web dream for decades now.
- triska 5y ago
  Yes indeed, although the semantic web chose a much more convoluted set of technologies without the syntactic and conceptual coherence and simplicity of actual logic programming languages, thus making reasoning about everything extremely complex and inconvenient in comparison. I think this is a severe handicap when using these technologies in practice. All automated reasoning is by necessity exclusively syntactic in nature since the meaning of processed symbols is inaccessible to a machine, therefore a careful consideration of syntactic aspects is very important to ensure easy analysis and meta-interpretation of logic programs and queries. As I see it, a good formalism for knowledge representation makes it easy to reason about the knowledge and also about all programs and queries that themselves reason about the knowledge, enabling easy meta-interpretation and coherent extension languages.
beaconstudios 5y ago
well if you think about it, a triple store is basically a database of propositional logic statements.
LukeEF 5y ago
I know at least one vendor of a prolog based graph database [1] that hopes the interest continues to grow. [1] https://github.com/terminusdb/terminusdb https://github.com/terminusdb/terminusdb

by triska 5y ago

sfvisser 5y ago

This has been the semantic web dream for decades now.

triska 5y ago
Yes indeed, although the semantic web chose a much more convoluted set of technologies without the syntactic and conceptual coherence and simplicity of actual logic programming languages, thus making reasoning about everything extremely complex and inconvenient in comparison. I think this is a severe handicap when using these technologies in practice. All automated reasoning is by necessity exclusively syntactic in nature since the meaning of processed symbols is inaccessible to a machine, therefore a careful consideration of syntactic aspects is very important to ensure easy analysis and meta-interpretation of logic programs and queries. As I see it, a good formalism for knowledge representation makes it easy to reason about the knowledge and also about all programs and queries that themselves reason about the knowledge, enabling easy meta-interpretation and coherent extension languages.

beaconstudios 5y ago

well if you think about it, a triple store is basically a database of propositional logic statements.

LukeEF 5y ago

I know at least one vendor of a prolog based graph database [1] that hopes the interest continues to grow. [1] https://github.com/terminusdb/terminusdb https://github.com/terminusdb/terminusdb