JCA is hardcore and does require alot of boilerplate that takes some getting used to and if all you need is a "connection pool", then the value of it is questionable when you can easily use something like apache commons pool to implement a quick easy solution. But when you start wanting to use transactions with that external system you're connecting to, even if they're only resource local, you start getting value out of JCA. I think JCA could be massively improved in a similar way EJBs improved from version 2 to 3. I haven't yet looked at the changes in JCA in jee7 but if i'm not mistaken, didn't David Blevins have alot of input into the new JCA spec?
Most people listed JMS and JPA (which require datastores) as parts of the stack they're using, so they're also using JCA under the hood right? -- View this message in context: http://tomee-openejb.979440.n4.nabble.com/Just-curious-what-parts-of-JEE-is-everyone-using-tp4673713p4673749.html Sent from the TomEE Users mailing list archive at Nabble.com.
