Re: [Neo4j] First-class type property on relationships but not nodes; why?
Am 05.05.2011 19:19, schrieb Rick Bullotta: In general, I think it's a good idea to avoid strings as types for a whole host of reasons (performance, future renaming/refactoring, etc.). Having type-ids for performance reasons instead of type-strings is okay, but because of future renaming/refactoring?!? What will you do when you have two seperate (semantic web) graphs on seperate machines and now one of the graphs moves to the other machine? Do you sync/renumber the type-ids of one of the graphs? Or do you support two Ids for the same type-string? I think this problem is today in the linked data area more important than supporting someone who just want's to change the name of some vertices/edge types. For this we could always use MapReduce (without the reduce of course ;). In the end having type-strings or better type-URIs is slow but the only real future-proof solution ;) Cheers... Achim ___ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user
Re: [Neo4j] GSoC 2011 Neo4j Geoprocessing
Am 30.04.2011 00:29, schrieb Andreas Wilhelm: I'm happy to say that I have been selected for the Google Summer of Code 2011 for my proposal Neo4j Geoprocessing. Hi! Nice to have someone on board for all this GIS stuff ;) The idea is to integrate a similar range of geospatial functions in Neo4j as PostGIS offers and implement a uDig plugin which uses the geoprocessing functions to analyze OpenStreetMap data. If you do so, take a look how to deal with all these energy lines, wind- and solar parks within OSM ;) Thanks so far... Achim ___ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user
Re: [Neo4j] Self-referencing relationships
Am 26.04.2011 18:20, schrieb Andreas Kollegger: What does you model look like, that you expect to require self-referencing? Oh, I think a loop is just the smallest circle within a graph. So as Neo4j supports graphs instead of just hierarchical trees it would be a natural fit ;) Perhaps it would be best for both sides of the discussion to support a compile time or runtime switch turning loops/selfedges on|off. Just my 2'edges ;) Achim ___ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user
Re: [Neo4j] .NET Bindings
Am 28.02.2011 10:13, schrieb Peter Neubauer: Both http://neo4jrestsharp.codeplex.com/ and https://bitbucket.org/2hdddg/neo4restnet/overview have not been worked on for a number of months. Hi... I think it depends. For normal operations http://neo4jrestsharp.codeplex.com might be best, although the other one is a bit more LINQy. I would suggest taking over one of them, base it on GIThub and start getting a few interested people involved. There is a project called Neo4Sharp I started in combination to my other projects like Blueprints.NET and Pipes.NET which is located at https://github.com/ahzf/Neo4Sharp Because of very special reasons there is currently no source code available, but I will release some in the next couple of days and weeks. The main goal of these projects is to provide a reimplementation of the Tinkerpop graph processing stack for .NET with additional benefits. Everything is still under heavy development, but Blueprints.NET and Pipes.NET are already quite useable. I'm still looking for some more people to contribute and to make use of them in real customer applications ;) Magnus MÃ¥rtensson would certainly be interested I guess. Also, there is a thought to use OData as a bindning prototcol to the .NET world, but I have no good idea on how well that approach would fit Neo4j? It seems however to be big in .NET land? OData is more semantic and a lot of conventions than syntax as there are multiple representations (JSON, ATOM, ...) for it. But yes it would be a very cool feature to access graph data via OData. Cheers... Achim ___ Neo4j mailing list User@lists.neo4j.org https://lists.neo4j.org/mailman/listinfo/user