Hi Balazs,
On Fri, 26 Apr 2013,
zsoldos.bala...@gmail.com wrote:
> What I am missing is more details about the project. Chapters like:
> * Roadmap with due dates: I do not miss exact dates but I see the same
> status page since half a year. I do not know when I can expect 1.0.0
> (not even the year) and if you suggest using this vm to be used in
> production.
You're right that the status page doesn't change very often, and the
latest release (0.6) is pretty old. I'm hoping to do a 0.7 release soon,
probably in the next month or two.
There's no concrete roadmap or timeline for 1.0. There are various
features I plan to work on in the next year or so, including invokedynamic
support, trace-based compilation, and value type optimization, but I don't
know whether they are pre-1.0 or post-1.0 features. And I'm not sure it
matters much; pretty much everyone involved in the project right now is
using a clone of the Github repo, so release numbers don't mean much to
us.
I do understand that there is value in publishing stable, well-tested
releases from time to time, so I'll continue making them, usually about
once a year. Are you hoping for more frequent releases? Is the Git repo
too unstable for your purposes?
Anyway, regardless of when and how often we release, there's no specific
roadmap to follow. To my knowledge, here hasn't been much interest in
organizing the project that way. Instead, we just work on stuff we find
interesting (or which our employers find interesting), sometimes
collaborating, sometimes individually.
> * Success stories: Information if this VM was used before in production
> environments and if there were what were the issues others faced?
The company I work for has been using Avian in production for many years
now. That's a special case, of course, since we wrote it from scratch,
and thus we're uniquely capable of fixing any issues that arise. So I'm
not sure our successes and setbacks are relevant.
I know of a few cases where it's used in production by other companies,
but I don't know if they necessarily want to advertise that fact.
Perhaps others on this list have experiences to share.
> * Dynamically updated benchmark results: As the first thing I did was that
> I wrote some really primitive applications to test the speed of Avian
> (only recursive function calls and iterations). I found that the Oracle
> JVM is about 2-3 times faster than Avian. It would be nice if there was
> a dinamically updated benchmark site where people could check the
> improvements of performance. For me around 50% loss of performance would
> still be a good figure against Oracle VM to release my OpenSource
> products in the way that they are tested on Avian as well.
Yes, I agree that would be great. Perhaps the DaCapo suite would be a
good place to start. Maybe we can make Travis CI run the suite on each
commit and compare it to HotSpot and few other VMs. There are probably
lots of cool things we could do with Travis. Too bad I'm too lazy to try
:)