Hi, On 1/16/07, Wolf Benz <[EMAIL PROTECTED]> wrote:
I sympathize with teh aspect for fast development but... I kinda still feel releaved with eth idea I can swap to JDO2 anytime I want. Untill I've seen JCR really deliver. (not being cynical - but it's just first time I use it) So I write against a DAO Interface as it allows me to quickly swap one implementation against the other. I'm developing a production app so I can't take any risks at the end of the day. (it's a already a miracle I can do this ;-)
That's a valid concern and a good reason to stick with DAO. I just wanted to point out this alternative. For a DAO architecture a (more or less) one-to-one mapping between node types and DAO classes is conceptually the best alternative.
That aside, hell yeah, I'd also like to see that taglib go official - why not for the 1.2 release already? The demand is hughe for such - believe me. I believe it's much more worthwhile to get that one out asap (better sooner than later) to allow everyone to contribute instead of all the little parallel devs going on now. (power to the people! :-)
We have quite a few structural changes in 1.2 so I wanted to keep the amount of extra pieces to a minimum to simplify things. PS. For a DAO architecture you might also want to check out the JCR Mapping component from the incubating Apache Graffito project (http://incubator.apache.org/graffito/). It's an object-content mapping tool for JCR like the many object-relational mapping tools for JDBC. Unfortunately, like the JCR tagblib, the JCR Mapping tool has not yet been released, so using it in a production application might be a bit risky. BR, Jukka Zitting
