I am interested in using OrientDB with the Ferma alternative to Tinkerpop2 Frames in a very large project, but I have had some concerns since Tinkerpop2 is no longer under active development. Tinkerpop3 is wholly uninteresting to me as it lacks the features I want (Blueprints, Frames or a Frames alternative like Ferma), it is too immature for enterprise-scale projects, and there are concerns about its performance.
On the Ferma Google Groups mailing list, the opinion was that Tinkerpop2 is unlikely to go away anytime soon (probably not in years) due to its widespread implementation in existing projects and the current state of Tinkerpop3.
I would like to hear the opinion of OrientDB developers and users on whether it still makes sense to base an enterprise scale project on Tinkerpop2 at this point in time. I believe I read in an earlier post that OrientDB plans on supporting Tinkerpop2 Blueprints as their graph API for at least 3 more releases (does this mean major or point releases?), is this correct? Ferma developers say that if Tinkerpop3 ever reaches a maturity and performance level comparable to Tinkerpop2, it will be ported.
From a practical perspective, OrientDB and Tinkerpop2 with Ferma definitely seems like our best choice, but we can't afford to lock ourselves into a technology that might have to be replaced anytime soon.
OrientDB will provide support for TinkerPop 3 in v3.0, so mid 2016. Today we have thousands of users and hundreds of clients that are running with current Graph API that is based on TinkerPop Blueprints 2.6. This means that the support for TP Blueprints 2.6 will be available for long time.