What I missed was how all of these have affected you - February's entry appears to be the only one about your usage, vs OrientDB itself?
Apr 9: I believe the context cannot be ignored. First @anonwarnings publicly says “read through their [OrientDB] bug tracker some time, you'll be scared”, then @lvca removes dozens of bug labels on Apr 8 and then the bug label color changes to light grey. It can be a coincidence but I doubt it.
Jun 5 & 16: common reasons for forking (and splitting a community) are a dead project or disagreements regarding architecture, feature set, etc. Oriento was still active and the only disagreement was the main contributor stating that OrientDB wasn’t the best multi-model database out there. The main contributor was not the only contributor, others had write access and there was an open path for the OrientDB team to push their changes (as they had done in the past). The thread quoted and linked on Jun 17 goes over more details. OrientDB was obviously in their right to fork but in my own opinion it was not the option that most benefits the community or even them.
After February we hit some small snags: queries that didn’t work as expected or small bugs with known workarounds. All stuff that we could find solutions for in stackoverlow, google groups or github issues and are not noteworthy. Our code still doesn’t work with OrientDB 2.1 RC-4, we suspect an undocumented breaking change somewhere and we haven’t bother investigating. Currently we are close to go live in production and essentially we are very nervous.
I don’t want to give the wrong idea, I still believe OrientDB has great ideas and it’s somewhat a flawed diamond. Unfortunately the flaws lie in the project management, developing methodology and testing practices. Unless these change I don’t see OrientDB becoming the rock solid database every database should be.
> The bugs are there yes. I hope they fix them soon enough because the product idea is very good.
I totally agree with you and 10 months ago I was really hopeful about v2 being the silver bullet. I doubt v2.1 will be it since it doesn't even run with our current code. Perhaps further ahead they will get there.