Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [jakarta.ee-spec] [jakarta.ee-community] Defining Jakarta EE 12 Scope in Program Plan

Scott (Red Hat): This is not really a valid comparison in my view. The OpenJDK project still is effectively a benevolent dictator model. I also don't believe the rapid cadence has been universally accepted.

Reza (Microsoft): I suspect we may need to agree to disagree on this one, but at least I can try to explain a bit more where we are coming from.

What we observe from our customers is that they indeed adopt Java SE versions very slowly on Azure (a majority is still on Java SE 8 and Java SE 11). Nonetheless, when it comes to evaluation-like activities they demand the latest Java SE version and say they are very excited about the newest features such as Records and Virtual Threads (and soon CRaC, native compilation, etc). This is the reason why on our services like Azure App Service we try very hard to stay up-to-date. We look at Jakarta EE the same way. The latest Jakarta EE Developer Survey appears to corroborate our view, including on faster cadence, Java SE version support, and adopting Java SE features sooner: https://blogs.eclipse.org/post/tatjana-obradovic/rising-momentum-enterprise-java-insights-2024-jakarta-ee-developer-survey.

We value vendor neutrality no doubt. It's one of the key reasons we are here. The problem is that our Java customers say what they actually care about far more is time-to-market. The vast majority of our customers Jakarta EE continues to lose to Spring Boot are particularly vocal on this point. Their eyes basically roll over the moment we try to talk about why vendor neutrality is something they should care more about.

This should logically bring us to another key consideration - does being multi-vendor have to mean being slower and having less to show for in releases? I have to say the Economist in me that has an active interest in Organizational Theory really does not think so. Effective alliances should be able to consistently out-compete most unilateral approaches. The fundamental reason for this typically is that good alliances can bring to bear greater collective resources, economies of scale, and strategic depth. They key is whether an alliance has a critical core mass of players that have fundamental alignment and an ability to collaborate effectively. I have always hoped that to be true for Jakarta EE and hope that it can materialize in a more compelling Jakarta EE 12 release.

On 10/27/2024 3:42 PM, Reza Rahman via jakarta.ee-spec wrote:

Apologies, Scott - I was actually just in the process of moving them. Thanks for just getting it done, saves me a bit of time. Let me try and respond one comment at a time to try to avoid overlong emails.

Scott (Red Hat): The deprecation model in Jakarta EE is still too conservative. As soon as a feature is deprecated all related TCK should be archived and no longer required for compatibility. Vendors are free to support the feature and run the archived tests, but they are no longer a requirement for compatibility.

Reza (Microsoft): Microsoft very much supports this view. I am not sure where Working Group current consensus is on this, but we would be happy to vote +1 on this where possible.

On 10/27/2024 3:17 PM, Scott Stark wrote:
Moving some comments I had made on the doc that I don't see here:

Regarding the Why is a timely, up-front deadline important? section, and the statement "but Java SE has done so in a far more effective way,"
This is not really a valid comparison in my view. The OpenJDK project still is effectively a benevolent dictator model. I also don't believe the rapid cadence has been universally accepted.

Regarding the Why does aggressive deprecation matter? section, The deprecation model in Jakarta EE is still too conservative. As soon as a feature is deprecated all related TCK should be archived and no longer required for compatibility. Vendors are free to support the feature and run the archived tests, but they are no longer a requirement for compatibility.

On Sun, Oct 27, 2024 at 12:46 PM Reza Rahman via jakarta.ee-community <jakarta.ee-community@xxxxxxxxxxx> wrote:
I am moving comments on my Jakarta EE 12 Google Doc
(https://docs.google.com/document/d/1U2qEqF9K969t5b3YuX4cwex5LJPvF3bt1w27cdKNpDM/edit?usp=sharing)
to Jakarta EE mailing lists when possible. The problem with Google Docs
comments is that they do not scale very well, aren't very readable on
smaller devices, and do not archive well. I will do so one email per
comment. The person commenting is copied.

Context: Why does replacing EJB matter?

Josh Juneau (Community): Are there any comprehensive tutorials on how to
utilize CDI rather than EJB for querying entities? It seems like these
tutorials need to be made front and center in an effort to help steer
people to CDI and to show that EJB is no longer needed in many cases.

Reza Rahman (Microsoft): Good point. As of Jakarta EE 11, it is indeed
possible to just use CDI now for basic CRUD in a transactional and
thread safe manner with Jakarta Persistence. The same for EJB
@Asynchronous and @Schedule. At the bare minimum, this is worthy of an
Eclipse Foundation newsletter article and/or JakartaOne talk. The
material could cover where EJB is not needed any more and where it is
still needed. The title could be something attention grabbing like -
"EJB is Dead, Long-Live CDI and Jakarta EE". We could also ensure a
revised Jakarta EE 11 Tutorial can avoid using EJB when possible. Maybe
Kito could comment on this? Additionally, the Marketing Committee has
been sponsoring some guides. Could we consider already starting an EJB
migration guide?

On 10/22/2024 5:30 AM, Reza Rahman wrote:
> Hi folks,
>
> I would like to see if we can define clear, compelling, and specific
> scope for Jakarta EE 12 as part of the Steering Committee Program
> Plan:
> https://docs.google.com/presentation/d/1xUNDHMP_qTHH1wA3m0yCmWVf_sHp41Qd7Opq3FhgINs/edit?usp=sharing.
> I believe this is of critical importance at this juncture. If I did
> not think so, I would not bother trying. I have detailed all the
> rationale here:
> https://docs.google.com/document/d/1U2qEqF9K969t5b3YuX4cwex5LJPvF3bt1w27cdKNpDM/edit?usp=sharing.
> For those that recall, something very similar was done for Jakarta EE
> 11, so this isn't exactly without precedent.
>
> I would like to see if this can be done in the following couple of
> weeks, when the Program Plan is due.
>
> Thanks,
>
> Reza
>
_______________________________________________
jakarta.ee-community mailing list
jakarta.ee-community@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/jakarta.ee-community

_______________________________________________
jakarta.ee-spec mailing list
jakarta.ee-spec@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/jakarta.ee-spec

Back to the top