OK, here are some of the use-cases:
1. In my model the type of a node is derived from the relations it has like in this post<http://blog.neo4j.org/2010/03/modeling-categories-in-graph-database.html>. So the first thing is to make sure we create both the node and its relation(s). 2. The same goes the other way around, when delete a sub-category in the example we want to remove both the link and the node in single tx. 3. There are application specific use-case, for example in our application we have "process network" that is a network represent a single process and we want to clone that network which means create an exact replica but with a different name. If I need to create the nodes and links using repeated calls to the RESTful API and for any reason I lose connection in the middle I end up with partially cloned network. I guess there are other use-cases but let's start with these. Ido. On Mon, Dec 20, 2010 at 10:10 AM, Jim Webber <[email protected]> wrote: > Hi Ido, > > At the moment each interaction with the RESTful service causes a back-end > transaction to occur. We have been hesitant to allow transactions outside of > the service boundary because they're an implementation detail, rather than > being part of the service contract. > > Could you tell us which use cases you're interested in? It may be that > extending the REST interface is feasible, rather than exposing transaction > demarcation facilities. > > JIm > _______________________________________________ > Neo4j mailing list > [email protected] > https://lists.neo4j.org/mailman/listinfo/user > _______________________________________________ Neo4j mailing list [email protected] https://lists.neo4j.org/mailman/listinfo/user
