Neo4j vs. Graph Found (ONgDB) (2020) | Hacker News Reader

Neo4j vs. Graph Found (ONgDB) (2020)

casetext.com

47 pointsotobrglez13 comments

There's nothing wrong with what Neo4J is doing here and they aren't fighting the GPL/AGPL. ONgDB forked it, with was fine, but they just couldn't resist trying to give the finger to Neo4J when they did it. In doing that they used Neo4J trademarks and made false or misleading statements and shouldn't be surprised that Neo4J would take issue with that. They're required to enforce their trademarks.

They should have just forked it, changed the names which it looks like that's what they're doing going forward.

Neo4j is so hostile to their open source "Community Edition" that I understand the sentiment that led to this.

It's severely crippled, practically untunable and unmonitorable. You basically can't run it in a production context unless you're willing to take huge risks that it can shit the bed on you at any time and you're willing to rebuild from scratch.

I'm in a position where I have to do this and I'd like to flip Neo4j the finger every single day.

Can you switch to the fork?

No. That fork is significantly behind the current codebase and wouldn't work with the client driver we're currently using.

Sorry it is falling down for you. I find that Neo4J community edition is perfectly functional - as long as you don't need to scale to multiple servers, then you have to pay.

re: untenable: can you give us some more details? JVM tuning? Neo4J specific?

TBO, I use RDF datastores about 10x my use of Neo4J, but I appreciate their free edition. Another vendor who handles free vs. commercial well (in my opinion) is GraphDB, where the free edition is scalable to the largest single server you want to run it on, but only allows 2 simultaneous connections.

Some un-solicited advice: if you really need to scale a large graph database, look at BlazeGraph that supports both property graphs and RDF. It is fast, injests data very quickly, and just a few servers run all of WikiData. That codebase became Amazon Neptune, BTW.

Not being able to have RBAC (or any users that aren't admin), not being able to use backup/restore, limited configuration of checkpoint windows, no query logging and no monitoring are all showstoppers for use in Production and completely gated behind an Enterprise license.

I'm not saying it's not functional, because clearly it is, but it has no place in a corporate setting.

I'm actually well aware of (much better) alternatives, but this tool was chosen by people in the org who gave no consideration to licensing, or requirements to run in production and are senior enough that there's no one to go around them to fix the situation.

Thanks for taking the time to answer my question.

This is from 2020, any updates on the development here?

Looks like they couldn't make it happen.

https://www.graphfoundation.org/blog/graph-foundation-resolu...

Layman's explanation?

IANAL and this is just my reading, but it sounds like Neo4j the company gave their enterprise database product away on an open source licence that allows forking, stopped doing this, got forked, and are now trying to pull that ability to fork back using trademark law of the sort you'd more normally see when a supermarket makes its own brand products look suspiciously similar to the branded product they're trying to rip off.

supermarket home brands don't often say "This is exactly Cap'n Crunch"

I remember seeing a brand of Sport drink in a grocery store (PowerRade maybe?) but it was clearly in Gatorade bottles as the logo was embossed in the plastic. (I assume they were being bottled in the same plant)

Don't bathrooms have vending machines that say "Compare with Boss by Hugo".

Reply on news.ycombinator.com