In general I would always choose less features over better quality. I
believe with many others that 3.0 was too big, 10 months now, (Dec 13th
2001) and should have be split in 2 at least. What is the overhead of
releasing every 6 months or 4 months especially when you already run a 8
months EAP? Are you doing so much testing and performance tuning that it is
not viable to release more often? the doc and even the OpenAPI could be
released more incrementally too.
By continuously releasing, bugs and performance problems could be fixed
faster providing a better experience for your users. Personally I won't
complain about the EAP quality (I think it is rather impressive to maintain
such level of quality while developing so fast such a gui intentive app) but
I like bleeding edge and I do not mind a few internal errors here and there.
However 95% of the other users of my company (we have close to 200 licenses)
are not migrating for months for fear of bugs in a new build. Faster release
cycle would guarantee that they get better software more often. More users
using your product means more meaningful feedback which in turn improve the
quality of your product. In addition I would love fast software all the time
not just after 6 months.
Pushing this to the extreme why don't you close more actively bugs and
performance problems as you encounter them and make a clean EAP every month
that get rid of most of them? Could be the rolling STABLE build, what a
Just my 2 cents obviously