Let's not mention JDK as a good example today because of: 1. almost every release was about the switch-case, and little improvement in Lambda. Now it looks like they do not have language architect although we know that Oracle has the architect. 2. Valhalla can be enabled in CLI but this produced incompatible bytecode, thus two versions in one JDK. And users fire a bug against Surefire. 3. dropped feature raw-string-literals, basically unstable. This is a new experience with Java because we trusted the Java that it would be backwards compatible with older versions.
So JDK started with an extreme and goes to another extreme. What changed was the transition from JCP to OSS. So it seems that nothing is so good with jdk and frequent releases. Simply the Java language is saturated. You won't find too many new features in far aways future. I understand that the top management wants to have frequent releases but the language is not capable. Another story is the JVM because the performance and the GC is always welcome to improve more frequently. On Wed, Nov 13, 2019 at 9:31 PM Elliotte Rusty Harold <elh...@ibiblio.org> wrote: > I'm hearing a lot of subtly different descriptions of what different > developers and subprojects use the term "milestone" for. > > It's certainly reasonable to have a major version bump in the plugin > and drop support for Maven 2.0. The goal that the plugin major version > should match the Maven major version does seem to be causing some > mishegas though. > > Even more problematic is the hope to finish everything on the wish > list for a plugin before declaring a 3.0.0 release. That can leave > users without a clearly supported, stable plugin for years. More > incremental releases that provide additional features and bug fixes > that are ready now would be very helpful. The JDK itself and the > Eclipse project are two big examples of projects that have stopped > pushing mega-do-everything releases in favor of more predictable, > time-based release cycles. I was skeptical when they announced the > change, but it seems to be working well. The Maven project might want > to consider whether something similar could work for us. > > -- > Elliotte Rusty Harold > elh...@ibiblio.org > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > For additional commands, e-mail: dev-h...@maven.apache.org > >
