Hi Nicolas, Thanks for your reply!
On Tue, Aug 06, 2019 at 11:24:39PM +0200, Nicolas Laurent wrote: > Heya Enrique, > > I have a big question about the project: what's the rationale for merging > parsing and the rule engine? > Do you derive some specific benefit from it? I would not call it merging, but separating. Parsimonius can be given any grammar and a production in that grammar to return a parse tree. And Syntreenet uses those parse trees independently of the grammar that describes them. So Syntreenet is totally independent of any grammar. The code in syntreenet only deals with the logic of a rule engine, and can be used in whatever language is more appropriate for your domain and can be provided by a PEG. Any rule engine will need to parse whatever is fed to it as rules and facts; Syntreenet externalizes that concern to the most powerful tool for the task. As an added bunus, the code is extremelly compact, at less than 1000 lines of Python code. So, in summary, separation of concerns, so that I can focus in Syntreenet on offering and showcasing a forward chaining algorithm that, I believe, can be more scalable than the alternatives, such as RETE :) Best, -- Enrique Pérez Arnaud
signature.asc
Description: PGP signature
_______________________________________________ PEG mailing list PEG@lists.csail.mit.edu https://lists.csail.mit.edu/mailman/listinfo/peg
