geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Struberg <strub...@yahoo.de>
Subject Re: Discuss: Move Geronimo to Attic
Date Wed, 08 Mar 2017 09:26:47 GMT
I see no lack of interest in Java EE to be honest. Of course Microservices are currently spilled
high on the hype cycle, but that will quickly blow up imo.
MS architecture is only very good for a certain kind of application. For most business apps
the granularity is way too fine grain and the missing TX handling is often a showstopper (even
if Managers don't see this yet).

You are certainly right that there is a lack of interest in the *huge* big-iron app servers!
So yes, TomEE, Meecrowave etc fill the sweet spot which is interesting for 85% of apps.

I also do not have a problem with the missing TCK. Of course it would be better to have one.
But the only real progress is currently in CDI and BVal and those TCKs are available under
ALv2 even.

The main problem imo is that the Geronimo server part is not actively maintained anymore and
OSGi is not a really good fit for JavaEE anyway. Not that OSGi itself is bad, but it's not
a good fit.
Don't get me wrong, the Geronimo AppServer was a big step 14 years ago, and all the people
involved in this effort back then layed a rock solid fundament for all that came after that.
But the architecture is still quite outdated imo and it didn't get maintained for way too
long.


Otoh there is really a lot of good technology available inside the geronimo project. 

* geronimo-jta
* javamail
* xbean (including finder, scanner etc)
* the specs
and quite a few other nice parts and they still get committs and love.

I'd definitly keep them alive. 

I'm aware that quite some older PMC members have historically been interested in the Geronimo
AppServer and not in maintaining the ee-commons part of the geronimo project. 
But instead of dumping the whole project I'd say we just retire the Geronimo AppServer and
consolidate and focus on the single pieces. There are potentially other things like Sirona-incubating
which we could move over as sub-projects even. 

Of course I perfectly understand if some of the older PMC members which are not interested
in the adopted roadmap want to retire.

txs for all the hard work!

LieGrue,
strub


> Am 07.03.2017 um 22:44 schrieb Alan Cabrera <list@toolazydogs.com>:
> 
> IMO, consultants and researchers are the earthworms of a vibrant OS community that meets
the standards sought after at the ASF.  I don’t see how we’re going to attract them. 
While the ideas posited on the mailing lists are pretty interesting, I just don’t see any
of the ideas attracting a larger active community.  The reasons for this are
> 	• the lack of interest in JEE
> 	• inability to use a reasonably current JEE TCK
> 	• the size and age of the legacy code base 
> 	• project members unable to commit time resources to mentor new members
> When one reads about JEE not being “dead yet”, one is actually reading about a very
small subset of the JEE spec.  To be sure, there are interesting problems still to be solved
within certain silos of JEE.  I can’t think of anything that would apply to the entire pantheon
of JEE bits; imo TomEE is already focused on the sweet spot of JEE bits that are still relevant.
 One is hard pressed to think of any JEE sub-system in Geronimo that is not already separate
project. The reality is that Geronimo was an amalgam of OSS projects and the industry has
preserved those JEE bits that are still relevant.  The "value add", in no small part, of Geronimo
was the comprehensive testing of the JEE pantheon in toto via the TCK.
> 
> Given that we cannot use a reasonably current JEE TCK, the project is prevented from
engaging in a role of JEE-commons of sorts.  Frankly, even if we were to get the current JEE
TCK, nobody really cares anymore and, as I mentioned above, the interesting JEE bits are already
being worked on elsewhere with their own specific TCKs.
> 
> The size and age of the codebase makes it virtually impenetrable.  When one precludes
spec commits, I think the last real commit has been about a half a decade ago; I wouldn’t
be surprised if it was longer.  I personally have been knee deep in it recently but find spelunking
through it very daunting.  I’d rather spend any free time I have in some greenfield endeavor.
> 
> I’m certain that other project members and passersby are of the same mind.  Since I
have such little time to do greenfield coding, I have even less time to mentor someone who
is interested in tinkering with the code base.  I’ve no doubt that others are of the same
mind on this as well; witness the dearth of replies to inquiries on this list.
> 
> There is a lot of blood, sweat, and tears in this project.  I, for one, am honored to
have been able to work with the world’s brightest coders on the planet.  I have a lot of
great memories, and hangovers, of our once vibrant community and it’s very hard for me to
start this thread.  I think we should shutdown.  If anyone had a real interest in any kind
of resurrection it would have happened by now.  
> 
> If we have consensus on moving to the Attic, I will start a vote.  How things are “wound
down” will be discussed in a separate thread, soon to follow. 
> 
> 
> Regards,
> Alan Cabrera
> V.P. Apache Geronimo
> 
> P.S.  Please resist the urge hop in and administer CPR.  Before jumping on the table,
be brutally honest and ask yourself if you are operating on actionable facts, or fond well
earned memories.
> 
> 


Mime
View raw message