[javaee-spec users] [jsr342-experts] Re: Java EE 7 roadmap

From: Werner Keil <>
Date: Fri, 31 Aug 2012 00:47:06 +0200


Thanks for your input, as to all the others before.
Sorry I mostly replied in the CDI thread, but a lot of it related to either
CDI or other annotations.

If certain ways to "sharpen" how e.g. WARs can do things currently the
domain of EAR can be done without breaking existing functionality and
features that is a good suggestion, but honestly, it feels more PaaS
friendly, so probably best to do some of these things right and carefully
(without the rush that introduced e.g. some of the bugs, SE7 currently
struggles with, though it has no real impact to EE in most cases;-)

Logging, I would personally prefer a real JSR for it if enough people like
you feel the need and pain. There are a few JSRs, Money probably another
good example, that cater to SE, but by their nature will have greater
impact in the Enterprise world. 310 another example. Take BeanValidation,
JSF or JDBC, you name it, they all use java.util.Date or Calendar.
Not all may be immediately changed to new types from these new JSRs, but
some sure will benefit. If JDBC or JPA work more towards non Relational
Databases, too, mapping or binding a Money type sure makes sense, too. With
limitations it also does for RDBs.

I'd see some of those like JSR 330 which via CDI became more important to
EE, than in most Desktop or Mobile use cases, too.
If you get a proper Logging JSR similar to what SLF4J does in many cases,
then this certainly won't make it into EE7, I highly doubt, it makes sense
in SE8 either, given without proper Modularity you can't throw out the old
stuff, same goes for 310, too btw;-)
However I am very confident, some of those make great sense in EE8,
especially with the PaaS focus, where an abstraction of logging ideally per
tenant would come handy.

At my current client we also have to do that, with more than one container
btw. Most of it by tweaking Log4J or other internally used loggers the
right way. So if an abstraction that works similar in different containers
existed, that would simplify not just how we provision our Private Cloud

Am 30.08.2012 21:43 schrieb "Linda DeMichiel" <>:

> Hi Antonio,
> On 8/30/2012 11:39 AM, Antonio Goncalves wrote:
>> Hi all,
>> I feel happy about this news and disapointed at the same time. Many of us
>> have expressed (at a very early stage) that
>> standardazing PaaS features in EE 7 was way too early. The JCP is full of
>> "standardizing early" gotchas and we felt it
>> would be another one. So I'm happy that this is postponed (and to be
>> honest, if it's not in EE 8 but in EE 9 it will not
>> hurt me). But I'm disapointed that we have spent so much energy in this
>> topic and left some behind. A few months ago I
>> mentioned the pain that developers have with logging. Will we be able to
>> solve it (like configuration, a single
>> container... and so on) ? It was clear to me that vendors where much more
>> interested in standardising PaaS features
>> (where there's money to be made) rather than wasting time on defining a
>> logging API (which is a day to day pain for
>> developers but doesn't have any business model behind).
> Now that we can focus better on the non-cloud aspects, I hope that we can
> make better progress here.
> As I said in my note, I'd like to see us getting going on at least some of
> these topics in advance of Java EE 8.
> I hope we will take this into account, like we should have done with
>> Entity CMP : standardizing too soon is a bad thing !
>> So let's keep the good work happening and, as you say Linda, focus on EE
>> 7 to make it follow the path of EE 6 :
>> enhancements in simplification and usability.
> thanks,
> -Linda
> See you soon on the mailing list
>> Antonio
>> On Thu, Aug 30, 2012 at 6:31 PM, Linda DeMichiel <
>> <mailto:linda.demichiel@**<>>>
>> wrote:
>> When we announced the Java EE 7 JSR back in early 2011, our plans were
>> that we would release it by Q4 2012. While this target date was three
>> years after the release of Java EE 6 and certainly later than we would
>> have liked, at the time it seemed like an aggressive schedule given
>> the proposed scope of the release. We have since adjusted this date
>> once (to the spring of 2013) in order to accommodate the inclusion of
>> additional JSRs of importance to the community (in particular, Web
>> Sockets and JSON-P).
>> As you know, our focus in the Java EE 7 release has been three-fold:
>> to continue to invest in significant enhancements in simplification,
>> usability, and functionality in updated versions of the JSRs that are
>> currently part of the platform; to introduce new JSRs that reflect
>> emerging needs in the community; and to add support for use in cloud
>> environments.
>> At this stage of the process, I think it is safe to say that
>> maintaining the entirety of this agenda -- particularly the aspects
>> related to PaaS enablement and multitenancy support -- puts our
>> proposed dates at very significant risk. We estimate that
>> realistically we would not be ready with a release of Java EE 7
>> until the spring of 2014. In our opinion, that is way too long.
>> After considerable soul-searching as to the causes of this delay --
>> limited industry experience in the cloud area when we started this
>> work, together with a lack of maturity in the space for provisioning,
>> multi-tenancy, elasticity, and the deployment of applications in the
>> cloud -- we are proposing that we defer to Java EE 8 the areas of PaaS
>> enablement and multitenancy support.
>> Of course, we continue to believe that Java EE is well-suited for use
>> in the cloud, although such use might not be quite ready for full
>> standardization. Even today, without Java EE 7, vendors such as
>> Oracle, Red Hat, IBM, and CloudBees have begun to offer the ability to
>> run Java EE applications in the cloud.
>> Postponing the remainder of the work on cloud support until Java EE 8
>> will therefore also have the important advantage of enabling Java EE
>> vendors to gain more experience with implementations in this area, and
>> will thus help us avoid risks entailed by trying to standardize in an
>> area that is arguably still some time away from being mature.
>> It is important to note that the features that we have already added
>> to Java EE 7 for cloud support -- such as resource definition
>> metadata, improved security configuration, JPA schema generation --
>> serve as enhancements to the Java EE 7 programming model in non-cloud
>> environments as well. The inclusion of these features in Java EE 7
>> will help expedite a cloud-oriented release of Java EE 8 in the
>> future.
>> We plan to target this Java EE 8 release for the spring of 2015. We
>> expect to include new JSRs for application configuration, for
>> JSON binding support, and others, which we hope to launch in advance
>> of the completion of Java EE 7.
>> This shift in the scope of Java EE 7 also allows us to better retain
>> our focus on enhancements in simplification and usability and to
>> deliver on schedule those features that have been most requested by
>> developers. These include the support for HTML 5 in the form of Web
>> Sockets and JSON-P; the simplified JMS APIs; improved Managed Bean
>> alignment, including transactional interceptors; the JAX-RS client
>> API; support for method-level validation; a much more comprehensive
>> expression language; and more.
>> To conclude, what we are proposing is to hold to the current dates for
>> Java EE 7 (spring of next year); maintain the focus on all of the
>> feature enhancements targeted at simplification and usability; retain
>> the cloud-related features we have already defined; and defer the
>> remaining portions of the cloud-oriented work to Java EE 8.
>> We feel strongly that this is the right thing to do, in view of what
>> we and our team have heard from members of the community.
>> Please let us know if you have any major concerns with this proposed
>> direction.
>> thanks!
>> -Linda
>> --
>> Antonio Goncalves
>> Software architect and Java Champion
>> Web site <http://www.antoniogoncalves.**org<>>
>> | Twitter <> | LinkedIn
>> <**agoncal<>>
>> | Paris JUG <> | Devoxx France <