My understanding of JMS is that it is up to the JMS provider to provide an implementation of the JMS interfaces defined by the Sun Community Process. In other words, an application Server such as ATG does not have to do anything whatsoever in order to be able to act as a JMS client - rather we obtain a JMS implenetation [jar file] for lets say IBM MQ Series, and place it in ATG Dynamo's classpath.
Is this a correct understanding?
The reason that I am a little unhappy with this understnding is that J2EE 1.2 did not require application servers to support JMS - Thus implicitly suggesting that indeed the application servers would have to do something in order to send messages to a JMS provider?
Can someone explain this please?
My understanding of JMS is that a program that uses JMS, whether in an application server or standalone, provides the code using the JMS APIs, and then uses third party jar files to provide the implementation.
I don't believe that the application server would have to do anything at all to send messages to JMS.
I thought that the new J2EE spec meant that J2EE 1.3 compliant application servers will have to provide an implementation of JMS.