Leaders of the JCP admitted this week that there are still kinks in JCP process (such as slowdowns in the launch of J2EE 1.4) but asserted that the JCP is still the best way to oversee Java standards. As a case in point, 4 Sun-submitted J2ME JSR's were voted down last month due to "memory footprint" and licensing concerns, but were later revised and then accepted.

Such is the process of evolution in the JCP.

Read Leaders Admit To Kinks In Java Community Process.

And from Cafe Au Lait:

"At the end of last month four Sun-submitted Java Specification Requests (JSRs) were voted down in the Java Community Process (JCP). These are

JSR-216, Personal Profile version 1.1 for the J2ME™ Platform
JSR-217, Personal Basis Profile version 1.1 for the J2ME Platform
JSR-218, J2ME™ Connected Device Configuration (CDC) 1.1
JSR-219, J2ME™ Foundation Profile 1.1
These four drafts would have updated the existing 1.0 versions of those specs to support Java 1.4. However, the nay-voters seemed seriously concerned about the increase in memory footprint this would have required.

Sun has now revised these JSRs to use less memory, and the revised JSRs have passed, though IBM and Motorola still voted no. There also appear to be some disagreements within the JCP board about the licensing of these specs. Sun apparently plans to charge a "Per unit royalty fee which will be dependent on volume."