You have a point there... I've seen people who see the pattern on the level with other things like goto statements. But I stand by my statement you quoted, because the flexibility directly correlates to the fact that they use EAV.

Businesses out there have so many different types of "products" and "attributes," and I've seen all of these packages attempt to tackle attributes... and I think the usage of EAV to tackle such a problem was justified.

That said, I definitely do not agree with their use of that pattern in some aspects of their software, though I can see why they did it. You can pretty much add fields to anything in Magento. Throw an extra text field here, and extra drop-down there. Makes it very easy for non-developers to add things.

The biggest problem is that it can cause resource issues and obviously makes writing raw queries a little more difficult... but I think the way it behaves with the code models is much cleaner than, say, having one single table with different columns for each data type.

I guess that's really up to argument though. Like I said, you bring up some good points.

Yay! Another book for an outdated version of Magento. At the rate we're going, Magento will never have apt documentation. I think the part that kills me is that if Varien had more of a penchant for providing public documentation for its community edition, more authors would be writing better niche books for Magento. Magento is so very flexible (the only free e-commerce system I've seen that uses EAV), and has great potential; however, only a paucity of developers know how to bend it to their will. The theming alone can be a bit of a learning curve, but development is outright in lacking in up-to-date documentation. And I don't blem the Zend Framework that's the holdup--it's learning to use their Mage extension of ZF on top of that. Gross resource usage aside, documentation is sparse, and that means subsequent books will take longer to write and publish, and that's not good for Magento at all I think. And I think it also affects the lack of the community stepping up to the plate. Most of the developers who went through hell understanding what's going on under the hood decide to sell their extensions for money, rather than release the source and documentation to the community at large. And hell, I can't blame them for having to spend all of that time without documentation--I'd be bitter as hell. Like I said, Magento has a lot of potential (and I mean it's already there--developers just need to learn to implement it). I think it has a lot of pluses that go beyond other popular e-commerce platforms (especially Ubercart in terms of the flexibility of attributes and extending beyond what comes out of the box), as well as A/B testing and the such. Having deployed sites from so many free e-commerce platforms, I think Magento is the best for enterprise-level implementation... but from my standpoint I still have to recommend it tongue-in-cheek because it always means far more development time due to so many unknowns.