activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From artnaseef <>
Subject Re: [DISCUSS} HornetQ & ActiveMQ's next generation
Date Fri, 27 Mar 2015 17:34:01 GMT
I agree with Hadrian again.  It's important for HornetQ to grow its own
community, and one that stretches beyond the Red Hat.  A project at Apache
only to satisfy the needs of a single entity does not seem appropriate.

*AMQ-6 Name*
I would also like to re-hash the concern of using the AMQ-6 name.  Doing so
raises issues.

Using the name AMQ-6 creates a presumption that HornetQ will whole-sale
replace ActiveMQ 5.x.  (Please understand I give 0 credit to the idea that
ActiveMQ 5.x could continue on as ActiveMQ 7.x - that's not reasonable).

While it is good to have a common direction and shared vision, this isn't
the right direction.  Using the AMQ-6 name puts users on notice that they
will be expected to switch; I know first hand - I started to worry about
Apollo and my team's efforts on 5.4 back-in-the-day, how much of that work
would need to be revisited, and whether Apollo would continue to meet my
needs or I would need to shop for an entirely new messaging solution. 
Which, by the way, proved to be a valid concern since Apollo didn't support
JMS in its earlier releases.

And there were missing-message and duplicate-message bugs I tracked down and
for which I personally contributed fixes; would those problems resurface
with a new solution?

It isn't wise for the ActiveMQ community to put all its users on notice that
they will be expected to switch to the new broker - at least not until we
have reached a point at which we, as a community, are convinced it is ready
to do so.  And that's a big "when" - I'm not sure it will ever happen.

*Details and Impacts*
Keep in mind that two solutions to the same problem never take the same
approach, and the detail can easily become overwhelming for users.  For
example, while it is very easy to install and start ActiveMQ for the first
time, setting up deployments and configuration management around ActiveMQ
takes time and effort.  Moving to a new broker that is a complete rewrite
means that details like these will turn into significant effort for users,
and there's no guarantee that it won't be a blocker for some of them.

So, while the strength of the code is important, it is not the only
consideration.  Programmers don't think about these things.  Developers do. 
We need to consider the entire solution, how it will be used, and the
impacts our decisions will have on others.

*My Vision*
What I would love to see happen here is HornetQ runs as a separate project
and folks who want to see it do well and take over the ActiveMQ market
contribute to it.  And, in due time, if it takes over the market, we all
have the opportunity to move and contribute with it, if we so desire.

At the same time, ActiveMQ continues on its own path, preferably with more
folks, from more places who continue to believe in it, have not tired of it,
and are investing in it.  Encouraging more committers, PMC members, and
contributors in general.

HornetQ can continue on its path of feature parity with ActiveMQ in order to
ease the transition for ActiveMQ users.  And, since ActiveMQ continues on
its own, there's less pressure on HornetQ to adopt every feature.

In summary, I feel strongly that it's the best path forward for HornetQ to
avoid the AMQ-6 name.  And I agree that building its own community and
learning the Apache way through the incubation process is valuable and the
best path forward.  I hope the HornetQ folks are comfortable to go through
the incubator.

With all of that said, I plan to continue to support ActiveMQ regardless of
the outcome.  In that light - anyone who wants to discuss ideas on how to
address open concerns, such as a lack of JMS 2.0 support, I welcome the
discussion; please raise those concerns in separate threads and let's start
the action down the path of resolution.


P.S. I hope to see folks at ApacheCon!  I'm going to talk about a vision and
ideas for fixing some of problems AMQ has had for a long time.

View this message in context:
Sent from the ActiveMQ - Dev mailing list archive at

View raw message