OPEN SPACE, EXPLAINED
No fixed speakers, no outside program. You and the room decide what two days are worth spending on — we just lock it in early so we're not building it live.
By the second deep-dive talk at a typical conference, your brain is fried. You nod through a couple more, and by the time you're back at your desk, most of it is gone — you never actually get around to trying any of it.
Here, the same conversation happens twice: once in a session, and again over dinner, on a walk, in the open afternoons. That second pass is what makes it stick — and it's the difference between meeting someone and knowing them. So when you're back at the office and want to ask a follow-up, you're messaging a friend, not a stranger who happened to be at the same talk.
Every JVM builder works in their own world. GraalVM's team doesn't see how Azul's customers use checkpoint/restore in production. Nobody outside a single vendor really knows whether a given enhancement solved someone else's problem, whether a build is actually a drop-in replacement, or what pain points you'll hit trying to switch.
Is CRIU worth the complexity, or is AOT with Leyden's fast warmup the better bet? What happens to GraalVM now that commercial support is winding down? Does the licensing even make sense? Is moving between OpenJDK distributions actually seamless, and how differently do vendors handle security patches? Does Java hold up in embedded environments? When does Alpine actually pay off for a JVM container, and when does it just cause glibc and DNS headaches? Nobody has all of these answers alone — that's what two days in one room with people from different JVMs is for.
Then there's the stuff that only surfaces once you compare notes across platforms. Does the OS actually give memory back to the JVM the same way on Linux, Windows, and macOS — or does that depend on the garbage collector and which madvise call it's using underneath? Do the hardware quirks one vendor tuned around even apply to your chips? Are the benchmarks everyone quotes measuring your workload, or just whichever number makes a release look good? And on the compliance side: does JTREG actually catch the bugs that matter, or is it theater? Does passing the TCK mean anything beyond a checkbox for procurement? Nobody sitting alone at one vendor gets to stress-test these questions against people who've hit them somewhere else.
There's no committee picking talks months in advance. About a week out, anyone with something worth discussing pitches it in a sentence or two, and we cluster the pitches into a schedule ahead of time — so day one starts with sessions, not with an hour of building a grid. It's still entirely built from what the room brings; we just do the building before we get there.
Arrival, then straight into sessions already locked in from the week before. Friday runs sessions through the afternoon; Saturday wraps by early afternoon, with people splitting off to network or travel back. See the full Schedule.
JIndia is part of the JUnconference family alongside JCrete, and it borrows shamelessly from the wider unconference world — especially the SoCraTes conferences and the events that inspired them.
SoCraTes BE, and its sister events in Germany, the Canaries, Chile, Crete, France, Italy, and the UK — plus SoCraTes Switzerland and SoCraTes Day Switzerland. Also CodeFreeze and I.T.A.K.E., and friendly conferences in the same spirit: OSCO, AGI'LILLE, and XP Days Benelux. We're not the first to try this format — just the first to try it for the JVM, in India.
Questionsjindia.org