Marcelo, On Fri, Jun 10, 2011 at 9:39 PM, Marcelo Barbudas <nos...@gmail.com> wrote: > >> No, >> It's not based on Gremlin a first experiment on lowering the learning >> curve to pattern searching for graphs. Feel free to give it a swing and >> report back! >> > > A couple more questions about cypher: > > * Is it supported via the REST api? We where not able to get it in in time, the plugin needs some love next week, it's in the works at https://github.com/neo4j/neo4j-cypher-plugin
> * How fast is it? Right now, it is based on Neo4j Pattern matching, so it's not fast in all situations. That is the point. As time goes, there can be different optimizers, possibly implementations using Gremlin, Scala adn K/V store to do stuff better. That's why we don't want it as a standard right now. That take a LOT more than just putting it out IMHO. > * Can it do all the stuff Gremlin is able to do here? Especially > collaborative filtering: No. It is intended for the 80% of "daily use cases", see http://docs.neo4j.org/chunked/snapshot/query-lang.html for the current things. Cypher is much like SparQL - declarative (WHAT), and Gremlin is another way of saying HOW instead of Java, Scala or JRuby. Cypher will not be turing complete and thus not be usable as a programming language. We are even contemplating of not letting it mutate in the first iterations, in order to enforce that notion. > 1 > http://www.slideshare.net/slidarko/problemsolving-using-graph-traversals-searching-scoring-ranking-and-recommendation Yep, that is the coolness that Gremlin rocks at! /peter _______________________________________________ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user
