Hacker News Reader

Comment by raganwald | Hacker News Reader

Database Analysts will tell you that many ORMs are themselves inner platforms, re-implementing constraints, views, joins and so on that should be encoded directly in the database.

View on HN

That has become my acid test for how good an ORM is.

Most fail horribly. I count myself lucky if I come across an ORM in some language that didn't visibly have joins bolted on the side in such a way that they just barely work, usually spending all the design value of the library in the project. (That is, merely by using one of these libraries you've often incurred technical debt on the spot.)

Just out of curiosity, have any passed this test?

Based on reading the documentation, I think Hibernate for Java would qualify. I've never used it, though, so I can't be sure. SQLAlchemy for Python also looks good, but I've also never had the opportunity to use it.

hibernate will make you throw up with all the verbosity and nonsense.

sqlalchemy, on the other hand, is good.

All sufficiently large CRUD apps either use an established ORM or re-invent them poorly.

That's not true, work with some legacy perl apps from the 90s, it's raw SQL all the way down.

ORMs aren't implemented in the database, so it's less of an "inner" and more of a "parallel" platform. I think this is an important distinction, since one of the reasons people use ORMs in the first place is better integration with the programming language of their choice. Moreover, there are often practical reason to re-implement those features you've mentioned.

Let's take constraints as an example. Can you use them to display a list of user-friendly validation messages when someone fills out a web form? Will they benefit from all the tools (version control, refactoring, static verification) your language of choice provides? How reusable will they be?

If there were databases with language-friendly APIs for these things, I bet developers would be less likely to re-implement them.

I was quite alarmed the first time I used NHibernate with C#. I had some very non descriptive error, and the only solution I could find was removing a constraint I had created in the db.

The flip side to that argument is that the "M" part of the ORM can be understood an artifact of the poor storage technology, and that the ORM+database system should really be understoood as an object store with transactions, not an API layer.

No need to re-invent OODBMS. The had their chance - and failed.

Reply on news.ycombinator.com