OracleBlog

Please review the following interesting article by Jonathan Holland.

Why Relational Databases end up being the bottleneck

For the appropriate context of this debate, you can review this forum discussion on Joel's web site as just one example.

Stored Proc to avoid frequent builds

As is evident from the title and content of my blog, I'm a database professional. Nevertheless, I try to stay as objective as possible when evaluating this debate. I think you'll agree that my earlier article on J2EE vs RDBMS was as dispassionate and even-handed as possible.

That being said, the author makes an excellent point that's valid on more levels than merely the most direct.

I'll concede that many people strongly object to leveraging the RDBMS based on ignorance. They program poorly, resulting in poor solutions, so they blame the technology that is so darn hard to understand.

Without a good, solid foundation in RDBMS I can certainly understand how every schema you design and all the PL/SQL you hatch-patch together is probably going to compare quite unfavourably to what you designed in your more familiar language.

I'll also concede that getting the necessary foundation in RDBMS to leverage its power and to design superior solutions (depending on the situation) requires a real investment in time and study.

My final concession is that, generally speaking, while slapping something together in .Net or Java with only the basics under your belt isn't likely going to result in an award-winning application, I'd consider it reasonable to expect that it would probably still wind up being superior than something you slapped together in PL/SQL with a similar domain unfamiliarity.

My point? To leverage the power of RDBMS you obviously need developers with good, solid understandings of RDBMS, and they can be very hard to find. On the other hand, if you do as much as possible in Java or .Net not only would it be easier to find experienced developers, but you could probably get away with a higher ratio of less experienced developers. That won't always get you the best end result, but it will save you time and money.

Of course, I'm not advocating avoiding RDBMS simply because it's harder to find experienced specialists. I've seen how quickly people can become very proficient with even advanced RDBMS like Oracle (by reading awesome blogs, for instance). Furthermore, I've seen the vastly superior solutions you can sometimes come up with when designing database applications after having made that investment.

So while it is valid to claim that the benefits of a technology don't outweigh the investment required to use it properly, it isn't valid to dismiss a technology as being inferior just because you didn't make that investment.

As always I've enabled comments and I'd love to hear your thoughts on Jonathan Holland's article, the debate on Joel's web site, my response, and any of your feelings on this topic. Thanks in advance.