> On Nov 7, 2025, at 4:43 AM, David Alayachew <[email protected]> wrote: > > Hello @valhalla-dev, > > I have spent the past week or so playing around with the JEP 401 EA, > gathering a bunch of data. I am nearing completion with the data-gathering > phase, and preparing myself to start writing my experience.
Good to hear, look forward to reading about it! > However, I wanted to confirm what type of experience report is wanted for > this JEP 401. It's been mentioned in multiple places that many of the > performance optimizations one would come to expect with Valhalla have yet to > be implemented yet. Understandable -- this is still just the EA of JEP 401. > > So does that mean that experience reports should avoid talking about > potential performance increases/decreases? Or should we focus on the > usability and ergonomics of the language feature? Or is this just a typical > experience report, where we just speak freely about what did and didn't work > when trying on a real world program? Oh, we're definitely interested in your experiences with performance. Let us know where you see value classes helping, where expected gains don't seem to materialize, or where there are regressions. As you say, there are some gains you might hope for that simply aren't implemented yet. JEP 401 is the first step on a long journey. I've summarized the notable limitations here: https://openjdk.org/projects/valhalla/value-objects#value-object-performance If you have a use case where addressing one of those limitations would really help, great, let us know about it, it's really valuable to have a real-world sense of relative significance. Just understand that it is not a goal of *this* JEP to address those limitations. If you see other problems, they may be implementation bugs, or they may indicate new problem spaces we haven't been focusing on. Either way, it's valuable to hear about them. These are things that are immediately actionable—either we should go fix the bug, or we should better understand and communicate the limitation.
