wow, it feels strange to have that discussion in 2025 after 20 years of practice :)
it's what we promote through Modello for years https://codehaus-plexus.github.io/modello/modello.html and what we do for each and every descriptor we write changing the practice for POM = the most sensitive place - may be risky we have many other descriptors to challenge the change: Modello itself, site next: how breaking is this? I suppose the impact is mostly for IDEs or editors, isn'it? I'd love to have feedback from tool maintainers if we do that change on POM, I'd love that we do the change on every descriptor that Maven ships: settings, toolchains, extensions, ... but let's discuss and test on others parts of Maven: archetype, build cache extension, site, assembly, ... we have many places to test and let time to decide for Maven itself On 2025/07/22 12:47:05 Guillaume Nodet wrote: > Hi all, > > I'm writing to initiate a discussion about an important decision regarding > Maven 4.0's namespace strategy. We have two competing approaches > represented by different pull requests, and the *primary question is about > timing and acceptable breakage* for this release cycle. > Background > > As we approach Maven 4.0 GA (currently at RC-4), we have two PRs that > represent different philosophies: > > - *PR #2475*: Create a new modelVersion 4.2.0 (my approach, targeting > 4.1.0/master) > - *PR #10952*: Keep Maven Namespace the same (@elharo's approach) > > The Core Issue: Timing vs. Technical Merit > > *My position:* I don't disagree with @elharo's proposed change from a > technical standpoint. The benefits are: > > - Adheres to XML namespace best practices > - Makes XML processing tooling easier by not having to convert between > namespaces > > *However*, this represents a *major breaking change*. While my PR targets > 4.1.0 (master branch), since we already have a stable branch for 4.0.0, the > timing question remains critical for our overall versioning strategy. > My Proposal: Defer to Maven 5.0 > > Instead of rushing this change into 4.0, I propose we: > > 1. *Continue with my approach for Maven 4.1.0* (new modelVersion 4.2.0, > already targeting master) > 2. *Plan properly for Maven 5.0* with a brand new namespace that we > would then consider stable > 3. *Design the Maven 5.0 namespace from the ground up* with long-term > stability in mind > > This approach would give us the benefits @elharo is seeking while allowing > proper planning, testing, and ecosystem preparation. > Key Questions for the Community > > 1. *Timing*: Should we introduce a major breaking change in the 4.x > series, or wait for 5.0? > 2. *Risk tolerance*: What level of ecosystem disruption is acceptable > for a 4.x release? > 3. *Long-term vision*: Would a properly planned Maven 5.0 namespace > overhaul better serve our goals? > > I'm interested in hearing the community's thoughts on this timing vs. > technical merit trade-off. > > Best regards, > Guillaume Nodet > ------------------------------ > > *References:* > PR #2475: https://github.com/apache/maven/pull/2475 > PR #10952: https://github.com/apache/maven/pull/10952 > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
