That is exactly my point, since the chance of finding a developer who knows how hibernate is working is rather low, i’d rather not use it, since this might be a future risk of the project. I know this can be theoratically argued, but in practice this is a factor you have to take into account. This of course applies to hibernate and can not be generalized.

Indeed – if you think a tool like Hibernate will completely hide your database – you are on the wrong track.
But still, Hibernate does a good job at helping us to manage our persistence needs and even though the database pops up back at some places, there are clean ways to handle the issues – provided you know how the tool is working.

Remember: a fool with a tool is still a fool…

We are working with Hibernate since versions 1.x and are very happy with it. We came to it because we quickly realized that EJB persistent entities wasn’t an option for us (at that time and in our context) and found Hibernate as a potential candidate. It wasn’t a “standard” yet so we had to convince people… You know what I mean: everybody is talking about EJB as *the* solution and you come and say it “sucks”, lets go for something else. Anyway, it was a wise decision and never had to regret it.

Oh well, it wasn’t fun all the time. The product was still very young and we hit some bugs and other sorts of shortcomings… We even had to dig into the code and provide fixes to the community. It was very funny and exciting. Today the application is still in production (5 or 6 years) and customer is still very happy with it.

As a result, we got a very good understanding on how Hibernate works, what is the rational behind it and what it is good for or not. It always makes me sad when I see people using it without actually understanding the product. How many times have I see someone calling save() every time she makes a single change to a persisted entity… How many times the application is performing badly because of the n+1 fetch problem, or the lazy loading issues, etc. And even transactions – most people don’t see the need for a proper transaction management!

It sometimes makes me cry: Oh God, is it possible so many people don’t even understand what they are doing ? How comes those people pretend they know the technology? It is enough to successfully persist a single entity to master the entire framework? Even worst: most of the time they are not even *aware* of the problems – and ultimately, if it doesn’t work or perform nicely, they blame the technology for not being good enough. Why don’t they first question themselves? The product is used by millions of people on thousands of different projects and strange enough, people declare it is not suitable for their own particular case. Are they sure they understand how to use it? They probably don’t even ask the question…

I remember someone who was having problems with a shell script he wrote in bash. Bash… you know, that thing exists since ages and is in place in nearly every unix operating systems. After 2 hours looking for the problem, he decided it was a bug in the bash interpreter itself! Well, the bug was actually on the other side of the keyboard…

A wise guy once said: “le 21ème siècle est le siècle de l’amateurisme”… Think about it – it has never been more true!