activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daniel Kulp <dk...@apache.org>
Subject Re: [DISCUSS} HornetQ & ActiveMQ's next generation
Date Thu, 26 Mar 2015 20:59:30 GMT

> On Mar 26, 2015, at 4:22 PM, Hadrian Zbarcea <hzbarcea@gmail.com> wrote:
> 
> I don't buy the premise. I could argue that the promise of Apollo hurt the evolution
of activemq 5, because everybody waited for something to happen there.

That’s certainly possible, but that doesn’t change where we are today.


> I could also argue that cxf should have been an axis2 subproject called axis3.

This situation is very different.  Not a single one of Axis2 developers thought that the proposed
CXF architecture was a better architecture and was a way forward to solve their own issues.
 Quite the opposite in fact.   They thought the CXF ideas were significantly worse and not
a good path forward.   Plus, the active Axis2 developers had full plans, a roadmap, active
feature development, etc… toward future versions of Axis2.     In addition, the CXF developers
had no plans (and still don’t) to help folks migrate from Axis(1/2) to CXF.  Neither side
of the fence really had any interest in the other side.   This is different in that we have
some folks from AMQ5 thinking that the new code is definitely the right thing to do, we have
the folks from the new codebase excited about getting it to be the way forward by working
on the enhancements needed to get it there.   I’d like to foster that excitement rather
than stifle it which I think this thread is beginning to do.


> I did buy the premise a week ago, and I would have said the same thing (actually I might
have said it). But now I don't. The funny thing is that nobody even tried to deny that the
activemq6 is very important for hornetq and, as it seems, perceived by its proponents as a
key ingredient to its success. And nobody in the activemq community was desperately shopping
for a new broker 6 mo ago out of fear that activemq will die. For that reason I consider the
activemq eulogies as disingenuous.

I agree about the eulogy thing.   I see no reason why releases of 6.x would have any impact
on possible enhancements and fixes to 5.x (if we get any).   CXF still supports 2.7 yet 3.1
is coming out soon.   Karaf is supporting and enhancing  2.x, 3.x, and working on 4.x.   
 Ideally, a user could use 5.x and 6.x together in a cluster/network, but we’re likely a
ways away from that. 


> If one pays attention to the users@ community, users don't ask for a "new" broker. There
were questions about the future of the v5 broker though. And there was an answer (iirc from
Gary) saying something like "expect that to get very stable". There are some features requested,
yes, contributions are also welcome.

And how does having a 6.x version in development change any of that?  If people have some
fixes or enhancements they need for 5.x, great!

Dan


> 
> So why is this all happening? Again?
> 
> I think it's a moot point now, anyway, because I believe the new board will take a look
at this and will provide some guidance. Personally, I don't think I have much to add.
> 
> Cheers,
> Hadrian
> 
> 
> On 03/26/2015 03:10 PM, Daniel Kulp wrote:
>> 
>>> On Mar 26, 2015, at 2:42 PM, Hadrian Zbarcea <hzbarcea@gmail.com> wrote:
>>> 
>>> Perfect, but that was not the initial promise. What you suggest, David, can very
well happen in the incubator.
>> 
>> Personally, I’d prefer it to be done here.   I completely agree with David’s
response.
>> 
>> In a separate community, it would be completely up to that community as to weather
any of the ActiveMQ migration ‘issues’ are important to address or not.   Doing it here
means that not only do we get to prioritize that, but we’d also have (hopefully) the expertise
to make sure those issues are addressed.  To me, that’s the best way to make sure the “next”
ActiveMQ is actually a better ActiveMQ.
>> 
>> I don’t see why it cannot be "released" as  6.0.0-M1.   We can keep doing M# versions
until it gets to a point where the community feels it really is a “reasonable drop in”
replacement for AMQ5.   If it never gets there, fine.  If it does, great!   If the community
decides that going with an “enhanced” AMQ5 based thing for 6.0, we start doing M# based
on that code until THAT is ready.   Until the “6.0.0” final release is done, we have a
lot of flexibility and control.
>> 
>> From my perspective:  do I see ANY of the current AMQ5 contributors willing to spend
the time and effort to move the AMQ5 codebase forward to a completely new and enhanced architecture
and such?  No.   It hasn’t happened in the last 4 years, I’m not sure why it would happen
now.  Except for Art, all of the most “active contributors” have pretty much stated that
the new AMQ6 codebase from the grant is the better way forward.   That says a lot to me. 
 That said, if they really believed that, I also would have expected some contributions from
those committers to the new code base and that hasn’t happened either.  So I DO have a concern
about that.
>> 
>> Anyway, those that are objecting to this being called “6.0.0-m1”:   what are
your proposals and thoughts about how the AMQ community can move forward?   Are you guys going
to take up all the new work like JMS2.0, core scalability, etc…?
>> 
>> And to those that like the idea of moving forward with the granted code:  are you
willing to start helping to add the missing features like the kahadb stuff and OSGi support
and basic web console and such?
>> 
>> Dan
>> 
>> 
>> 
>>> 
>>> The way it's done right now is actually a very hostile takeover.
>>> 
>>> Hadrian
>>> 
>>> 
>>> On 03/26/2015 01:12 PM, David Jencks wrote:
>>>> I'm baffled.  I have (unfortunately, wish I had more time) very marginal
involvement with activmq these days and it was obvious to me (even if wrong) that replacing
the broker was the only plausible reason to bring in hornetQ code.  So if that is the intention
the obvious integration strategy to me is to start with the new broker code and add in all
the non-broker bits from activemq 5.  Isn't this what has been happening?  What other possible
integration strategy is there?  I said it before but I'll say it again,  I really don't understand
why everyone here isn't saying, wow, we just got a new broker and some new committers who
have the skills to write a broker, this is wonderful, how many years of work does that save
us, let's all pitch in and make sure it has all the features of activemq 5 and is as compatible
as we can make it.
>>>> 
>>>> thanks
>>>> david jencks
>>>> 
>>>> On Mar 26, 2015, at 12:29 PM, artnaseef <art@artnaseef.com> wrote:
>>>> 
>>>>>> 5.x needs a new core.
>>>>> 
>>>>> I think this point is really at the heart of the entire disagreement
here.
>>>>> 
>>>>> The initial grant vote did not mention that HornetQ was going to be taken
as
>>>>> a *replacement* for the entirety of ActiveMQ.  As several folks have
>>>>> mentioned here, we had the impression the code was going to be made
>>>>> available for merging into the ActiveMQ code base.
>>>>> 
>>>>> If the initial vote had been, "[VOTE] accept HornetQ as ActiveMQ 6 to
>>>>> replace the existing code base", the results of the vote would have been
>>>>> very different.  It may still have passed, but there would have been
this
>>>>> same discussion back then before heading part-way down this path, and
there
>>>>> would be no reason to discuss it now.
>>>>> 
>>>>> Chris - I think you mentioned there was a vote to bring HornetQ folks
into
>>>>> the AMQ PMC.  I don't believe that happened (someone please correct me
if I
>>>>> have it wrong).
>>>>> 
>>>>> 
>>>>> 
>>>>> --
>>>>> View this message in context: http://activemq.2283324.n4.nabble.com/DISCUSS-HornetQ-ActiveMQ-s-next-generation-tp4693781p4693856.html
>>>>> Sent from the ActiveMQ - Dev mailing list archive at Nabble.com.
>>>> 
>> 

-- 
Daniel Kulp
dkulp@apache.org - http://dankulp.com/blog
Talend Community Coder - http://coders.talend.com


Mime
View raw message