Now that Beehive is formally in Beta, it's time to begin thinking
about the road towards producing a 1.0 release.  It's vitally
important that as a project we don't just continuously churn and add
new features, but that we also deliver a usable, useful, and stable
platform upon which  real applications can be built.

The best path (IMO) to 1.0 is primarily based upon tracking and
resolving issues, and this takes effort from both the community of
users and the community of developers involved in Beehive:

- for users, please, please, please download the Beta distribution and
use it (http://incubator.apache.org/beehive/downloads.html).  Samples
are a great starting point, but use it to build something real.  
Please open a JIRA issue for anything and everything that stands in
your way... bugs, inadequate documentation, missing features,
whatever.   The only way we can know if we are there (at 1.0) or not
is if the user community helps us!     A key field for new JIRA issues
is the "Priority" field.   Big pain points for usefulness and
usability should be classified as  Blocker, Critical, or Major.  
Annoyances can be Minor or Trivial.  No issue is too big or too
small... we are looking for anything and everything!

- for committers and other interested parties, it is really important
that we stay on top of JIRA issues.   Please look for incoming JIRA
issues, evaluate, and take ownership of the ones in your area as
appropriate.   The key field here is "Fix Version".   Anything flagged
as "V1" is an issue we feel like we need to resolve prior to calling a
1.0 release done without jeopardizing overall stability.   Anything
flagged as "Unknown" (the default) implies that it hasn't gotten
attention yet and needs a domain expert to give it a look.   For bugs
that are assigned for V1, please stay on top of them as best you can,
because...

The next logical question is when is the V1 release?  In part the
answer will lie in what happens in the two areas above over the next
few weeks.   The difference in rates between incoming V1 issues and
resolution of those that have been identified will tell us a lot about
whether we are converging on the goal or diverging...  but the key
metrics will be the total number of bugs flagged as "Fix by: V1" (ones
we are seeking to close before release) and "Fix by: Unscheduled"
(ones that need triage and attention from domain experts).

I make these suggestions with no authority other than as someone
interested in Beehive being successful and useful, and in my past
experience the only way we can do this is to flush out the issues,
track them, and deal with them.   If someone has experience or
perspective on how to approach the "getting to 1.0 release" process in
a better or slightly different way, please speak up!

Cheers!

-- Kyle

Reply via email to