Jonathan Ellithorpe brought something to our attention on the user list https://groups.google.com/d/msg/gremlin-users/Ob2ScxBCcak/f-yc52Z-AwAJ
We have two features for describing Meta Property support: Graph.Features.VertexPropertyFeatures.FEATURE_ADD_PROPERTY Graph.Features.VertexFeatures.FEATURE_META_PROPERTIES I believe that ADD_PROPERTY was first and then META_PROPERTIES became the dominant term and it took over. ADD_PROPERTY as the feature would be symmetrical with VertexFeatures.ADD_VERTICES and EdgeFeatures.ADD_EDGES, but we sorta know meta properties as, well, "meta properties'. I sorta think we should stick with FEATURE_META_PROPERTIES as it is more widely used in the tests and deprecate ADD_PROPERTY. I can even add a test to validate that providers set them to the same value (until we can ultimately remove ADD_PROPERTY). In that same post he also noted that the SIMPLE feature category was not as "simple" as it could be thereby cutting off a reasonable number of tests that don't support mutli-property and don't actually use multi-property in their testing process. I'd propose that we we just create a FeatureRequirementSet.SIMPLE_WITH_MULTI to open up some extra tests in this area to providers who don't support multi-property features. I'll assume these changes are "breaking" as they include new tests that could disrupt provider implementations (however they should not create compile errors anywhere). If there are no objections in the next 72 hours (Tuesday, November 24, 2015, 6:00am EST), I'll assume lazy consensus and move forward with some JIRA tickets to get these changes in place.
