activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris Mattmann <mattm...@apache.org>
Subject Re: [DISCUSS} HornetQ & ActiveMQ's next generation
Date Thu, 09 Apr 2015 04:50:22 GMT
Hi Guillaume,


-----Original Message-----
From: Guillaume Nodet <gnodet@apache.org>
Reply-To: <dev@activemq.apache.org>
Date: Wednesday, April 8, 2015 at 3:44 PM
To: <dev@activemq.apache.org>
Subject: Re: [DISCUSS} HornetQ & ActiveMQ's next generation

>>[..snip..]
>>
>> The talk about no one wanting to develop ActiveMQ 5 isn’t backed
>> by the data. It doesn’t support that.
>
>
>> From reporter.apache.org:
>>
>> The project is frequently releasing code:
>> from: http://www.apache.org/dist/activemq/
>>   (5.11.0 was released in Feb 2015)
>>   (5.10.2 was released in Feb 2015)
>> from: http://archive.apache.org/dist/activemq/
>>   (5.10.1 was released in Jan 2015)
>>   (5.10.0 was released in Jane 2014)
>>   (4.9.1 was released in April 2014)
>>
>> 256 JIRA tickets created and 160 closed in last 3 months.
>>
>
>Lies, damned lies and statistics... ;-)
>Did I ever say that there was no development at all in the ActiveMQ
>project
>?
>
>That's not what I said. I talked about evolving the activemq 5 *core
>broker*
>code into a next generation broker for activemq 6, not about maintaining
>the activemq 5 broker and also not about adding additional protocols,
>persistence mechanism, etc...and certainly not about fixing bugs.
>When I write "*core broker*" explicitly, it rules out any non core
>components, which are more easily ported to a new broker implementation
>(see earlier points made by David Jencks).

And this is the precise point I’m trying to make. You can couch these
things
in architectural components all you want, but the fact of the matter is
that Apache projects are more than their singular components. You can
claim that
you were only talking about the “core broker”, and that’s the thing that
would be “soon [sic] abandoned for a lack of committers working on the
core 
broker.” The thing is whether it’s the ASF or not what you will find
developing
open source code for a long time is that certain parts of the code are
abandoned.
That happens in a long living project. New parts come in. The abandoned
parts
are picked by up again. This is the nature of software. Knowing this is the
nature of software is the reason that there is a PMC and not a “set of
software
developers” in each committee at the ASF. The PMC’s job and role is to
steward
the software - the board doesn’t care about things like parts of the code
being
abandoned and dying, etc. - these are reported each month in board reports
(
ideally). The board cares about wholesale imports of codes from external
sources that have the perceived impact of abusing names, community
goodwill,
and appear to be dominated by a lack of diversity on the PMC.

However, this is much more than simply a certain part of the code being
abandoned.
You and others continue to make the “crutch” that without the HornetQ
donation
that Apache ActiveMQ’s contributors would dry up and the project would be
at risk [implied]. My point in presenting the metrics to you and others is
that I highly doubt in my own long running experience at the ASF that a
project
that has been releasing as actively and with such activity would suddenly
dry
up on that activity. In fact, I know if the PMC was doing its job it
wouldn’t.
For example, releases are typically necessitated either by community
want/desire
and/or developer itch to scratch. If it’s the community end then there
appears
to be community desire around more than simply the broker that is in
question;
aka there are other parts of ActiveMQ; if it’s developer itch/scratch then
similarly the broker appears to not be the only reason that ActiveMQ is
released since there have been so many recently.

>
>So we're talking about the next major version of the *core broker*.  The
>attempt to write such a new broker was Apollo and it started in February
>2009.  This kinda implies the *core broker* is mostly in maintenance mode
>since a few years.

But the thing is - there is more to ActiveMQ than the core broker.
Otherwise
there wouldn’t have been a release in those few years, no?

Cheers,
Chris


>
>I'm not sure how the situation would have recently changed and that people
>suddenly want to start writing a new broker now, but I'm certainly wrong
>as
>I can't back feelings and experience by shiny statistics.
>
>
>
>> Cheers,
>> Chris
>>
>>
>>
>> >
>> >
>> >Cheers,
>> >Chris
>> >
>> >
>> >
>> >>So pointing to RedHat for abusive trademarks use or violation simply
>>does
>> >>not make any sense to me.
>> >>So I don't see that there is anything to fix, but clarifying how the
>>code
>> >>currently located in the activemq6 git repository  will be named,
>>either
>> >>activemq6 or something else, which can't be Apache HornetQ at this
>>point.
>> >>
>> >>
>> >>2015-04-08 18:09 GMT+02:00 Chris Mattmann <mattmann@apache.org>:
>> >>
>> >>> Hi Gary,
>> >>>
>> >>> Thanks. Well, we have a major problem then - see the
>> >>> subject of this email thread, and much of the discussion
>> >>> the last month. The discussion is one of these options:
>> >>>
>> >>> 1. Apache ActiveMQ has multiple products with multiple versions:
>> >>>  a.  ActiveMQ - (version 5.x.x)
>> >>>  b.  HornetQ - (which some are trying to call ActiveMQ version
>>6.x.x)
>> >>>
>> >>> 2. Apache ActiveMQ has 1 product with multiple versions:
>> >>>   a. ActiveMQ (version 5.x.x and version 6.x.x)
>> >>>   <—there is NOTHING in this option that mandates the current
>>HornetQ
>> >>> code becoming 6.x.x of ActiveMQ; also NOTHING stopping that.
>>Decision
>> >>> needs to be made.
>> >>>
>> >>> 3. Whatever is in the code repo now as ActiveMQ 6.x.x becomes
>> >>> Apache HornetQ (incubating)
>> >>>
>> >>> It sounds like you are taking 1b; and and 2a off the table. You
>> >>> are doing so, b/c Apache doesn’t accept code donations that are
>> >>> centered around names and trademarks that we don’t own; otherwise
>> >>> the product is renamed - the proposed renaming of it centers around
>> >>> abuse of trademarks since the proposed rename leverages an existing
>> >>> Apache product name. There hasn’t been work here to deal with the
>>due
>> >>> diligence of trademarks related to the HornetQ name.
>> >>>
>> >>> The community will need to have a plan for fixing that in its
>> >>> board report. I suggest working on that plan, rather than trying
>> >>> to correct my understanding. I also strongly suggest the community
>> >>> engage with trademarks@ and achieve something acceptable as I’m
>> >>> fairly sure that this isn’t.
>> >>>
>> >>> Cheers,
>> >>> Chris
>> >>>
>> >>>
>> >>> -----Original Message-----
>> >>> From: Gary Tully <gary.tully@gmail.com>
>> >>> Reply-To: <dev@activemq.apache.org>
>> >>> Date: Wednesday, April 8, 2015 at 8:56 AM
>> >>> To: "dev@activemq.apache.org" <dev@activemq.apache.org>
>> >>> Cc: <board@apache.org>
>> >>> Subject: Re: [DISCUSS} HornetQ & ActiveMQ's next generation
>> >>>
>> >>> >Hi Chris,
>> >>> >on 1) there may be a misunderstanding here. The code grant is just
>> >>> >that, code. there is no trademark grant. There is no intention of
>> >>> >having apache hornetq, that is not an option with the code grant
>>that
>> >>> >we have. Part of ip clearance and cleanup was to remove all
>>references
>> >>> >to hornetq. 2(3) was the intent.
>> >>> >
>> >>> >cheers,
>> >>> >Gary.
>> >>> >
>> >>> >On 8 April 2015 at 15:46, Chris Mattmann <mattmann@apache.org>
>>wrote:
>> >>> >> Hi Everyone,
>> >>> >>
>> >>> >> These are my following concerns as an ASF director that the
>>ActiveMQ
>> >>> >> community needs to address.
>> >>> >>
>> >>> >> 1. RH has a product, called HornetQ, which includes a website;
>> >>> >> branding, etc.  http://hornetq.jboss.org/
>> >>> >>
>> >>> >> At a minimum this is an extreme branding confusion if this
is
>> >>> >> ActiveMQ 6 and even more so if there is a HornetQ branch in
an
>> >>> >> Apache code repo. We don’t allow companies to come into Apache
>>and
>> >>> >> create confusion by importing their *still existing* products
>>into
>> >>> >> our neutral zone at the ASF and then keep maintaining their
>>external
>> >>> >> websites and so forth. This needs to be rectified, ASAP.  If
>>HornetQ
>> >>> >> exists in an Apache repo (which it does right now) -
>> >>>hornetq.jboss.org <http://hornetq.jboss.org>
>> >>> >> needs to go away at a date identified by the PMC in its next
>>board
>> >>> >> report.
>> >>> >>
>> >>> >>
>> >>> >> 2. The ActiveMQ PMC needs to deliver a plan for: (1) keeping
the
>> >>> >>community
>> >>> >> as 1 project with multiple “products”; or (2) spinning
out
>>HornetQ
>> >>>into
>> >>> >> Incubator or straight to TLP; or (3) keeping the community
as 1
>> >>>project
>> >>> >> with a single “product”. These are the only options. A
choice
>>must
>> >>> >> be identified and made by the PMC in its next board report.
>> >>> >>
>> >>> >> I would strongly encourage the community also to think about
the
>> >>> >> role of the PMC chair in all of this. To that point, the current
>> >>> >> chair has been the chair for *many* years and based on the
>>current
>> >>> >> status and issues in the community, I would strongly suggest
>>having
>> >>> >> a plan for potentially replacing the chair of the project.
It’s a
>> >>> >> healthy thing to do and these community issues may be better
>> >>> >> identified by some fresh blood and energy. I fully expect the
>>above
>> >>> >> issues to be discussed, and identified between now and April
22
>> >>> >> which is the next board meeting and the PMC’s report.
>> >>> >>
>> >>> >> Cheers,
>> >>> >> Chris
>> >>> >>
>> >>> >>
>> >>> >>
>> >>> >> -----Original Message-----
>> >>> >> From: Greg Stein <gstein@gmail.com>
>> >>> >> Reply-To: <dev@activemq.apache.org>
>> >>> >> Date: Monday, March 30, 2015 at 5:03 PM
>> >>> >> To: <board@apache.org>, ActiveMQ-Developers
>> >>><dev@activemq.apache.org>
>> >>> >> Subject: Re: [DISCUSS} HornetQ & ActiveMQ's next generation
>> >>> >>
>> >>> >>>Please note: earlier messages Jim has sent were as "Jim,
the
>> >>>individual"
>> >>> >>>using his years of experience at Apache to review the situation,
>>and
>> >>>to
>> >>> >>>provide feedback. Chris Mattman has also been assisting
lately;
>> >>>again,
>> >>> >>>as
>> >>> >>>"Chris, the individual".
>> >>> >>>
>> >>> >>>This message below is on **behalf of the Board**. Jim may
have
>>been
>> >>>the
>> >>> >>>messenger, but what is happening in Apache ActiveMQ is now
a
>> >>>specific
>> >>> >>>concern of the Board. As such, it needs to be addressed
per Jim's
>> >>>note.
>> >>> >>>
>> >>> >>>Regards,
>> >>> >>>Greg Stein
>> >>> >>>ASF Director, and Vice Chairman
>> >>> >>>
>> >>> >>>On Mon, Mar 30, 2015 at 7:53 AM, Jim Jagielski <jim@jagunet.com>
>> >>>wrote:
>> >>> >>>
>> >>> >>>> I think it has become somewhat obvious that this is
an issue
>>which
>> >>> >>>> is currently, as well as potentially, damaging an important
>> >>> >>>> project and community.
>> >>> >>>>
>> >>> >>>> As such:
>> >>> >>>>
>> >>> >>>> The ASF board has taken notice of some PMC issues going
on in
>> >>> >>>>ActiveMQ.
>> >>> >>>> There appear to be two distinct factions under the
same ASF
>> >>>umbrella
>> >>> >>>>of
>> >>> >>>> this project: One is focusing on a codebase called
"HornetQ";
>> >>>another
>> >>> >>>>is
>> >>> >>>> the more traditional Apache ActiveMQ PMC. What appeared
to
>>start
>> >>>off
>> >>> >>>> as an opportunity for these 2 factions to merge and
grow
>>together,
>> >>>has
>> >>> >>>> instead devolved into, for lack of a better term, a
power
>> >>>struggle.
>> >>> >>>>
>> >>> >>>> The board is not happy about the current state of affairs.
The
>>job
>> >>>of
>> >>> >>>> the VP is to be the eyes and ears of the board, and
clearly
>>that
>> >>>job
>> >>> >>>>is
>> >>> >>>> not being done effectively. The board offers its help
and
>>strongly
>> >>> >>>> encourages the PMC and the Chair to take it, before
more
>> >>>pro-active
>> >>> >>>> action is required by the board.
>> >>> >>>>
>> >>> >>>> We (the board) expect a full report by the current
ActiveMQ PMC
>> >>>and a
>> >>> >>>> roadmap for going forward, either as a single harmonious
>>project,
>> >>> >>>> or as 2 distinct projects.
>> >>> >>>>
>> >>> >>>>...
>> >>> >>
>> >>> >>
>> >>>
>> >>>
>> >>>
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>>
>>
>>



Mime
View raw message