Keep any evangelists away from these discussions. You don't need a salesman on your team.
DADom wrote:So much so that he asserted, if they’d been invented, Ralph Kimball would never have become involved with dimensional modelling and would have used graph databases instead.
My strength is identifying relationships too, so I prefer to store relational data in a RDBMS in a structure that is at least in BCNF. It is a proven way which guarantees consistency and correctness.
DADom wrote:Would be interested in you views.
According to the wiki, a graph db excells where data is "difficult to model" (and I agree with the comment "according to whom"). So, to evaluate it, you come looking for anecdotes.
I suggest you make a list (and check it twice) with advantages and disadvantages of both, and decide which are more important to you. THAT is the correct way out of these discussions - you will need to justify your choice, which is not done by pointing to an anecdote on codeproject, but by showing which of the two choices is the better fit in your specific case.
Personally, I'd say that the choice is obvious; Most RDBMSs are stable and well-optimized for data storage and retrieval, and there are a lot of resources on SQL92 (talk about backward compatibility eh?) Graph-databases would require something else than SQL probably, adding to the costs (and riscs).
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.