We have a “chicken and egg” problem here. We can’t see how something bakes/matures in the market without shipping it in products. So I think there is a middle ground
option here before making it mandatory for all in a platform profile.
First Jakarta Data goes GA on 1.0.
Second a number of vendors commit to shipping a Jakarta Data implementation in their products on the Jakarta EE 11 timeframe.
Third adoption and usage is reviewed to see if there is merit to inclusion in a platform profile.
I think this can give real world feedback as to the maturity of the spec in the real world.
Same comment for all 3 under discussion.
--
|
Steve
Millidge
Founder & CEO
|
Hello Platform Dev,
I have attempted to capture prior discussion about this topic from the following sources:
1.
Agendas: Previous agenda google docs
o
2023
§
Monthly
§
Regular
o
2022
o
2021
o
2020
2.
GitHub Issue: The corresponding issue in
https://github.com/jakartaee/jakartaee-platform/issues.
3.
Platform-dev Archive: Any mails on the archive at
https://accounts.eclipse.org/mailing-list/jakartaee-platform-dev.
1.
Agendas
2023:
·
Jakarta Data
o
Monitoring the progress
o
Too early to decide now
o
TCK isn’t written yet
o
Will run a formal vote on the mailing list to decide whether to include or not later
o
Target profile start: web?
§
Inclusion history
·
First time at the gate
·
Spring Data similarity. Layer on top of JPA.
§
Data does define a number of aspects of platform integration.
§
Jan shared some aspects of Spring Data vs. the way Jakarta Data is proceeding.
·
There was some disagreement about the extent to which the Data team did reach compromise.
2.
GitHub Issue
https://github.com/jakartaee/jakartaee-platform/issues/640
Hantsy wrote:
I have used Microaunt Data and Spring Data in projects, compared to these existing Repository solution, I think Jakarta Data needs more time to enter
the Jakarta main release train. Currently in Jakarta EE 10, we lack reactivestreams(or Java 9 Flow) and context propagate support in the core specs, such as http, cdi, tx, etc. Jakarta EE 11 discussion leaves some room for the common spec, we could consider
these. Personally I like the idea, and I would like to use all akarta EE specs seamlessly like using Spring in the application development, not spend time researching the gap between specs. I am not sure if the Jakarta Data 1.0 will include async, and reactive
streams( or Java 9 Flow API) support and if they(eg. Java 9 Flow) are supported in other spec(servlet, faces, rest).
Scott Stark wrote:
Red Hat will vote no on its inclusion. It needs more bake time in implementations.
Werner Keil wrote:
@starksm64 Hopefully
Red Hat also finds the time to participate as other provider reps do, especially given its mother company IBM co-chairs the spec? ;-)
Whenever it's considered ready (hopefully by Jakarta EE 12) I would not outrule the Core Platform either, but at least Web sounds fine.
Nathan Rauh wrote:
For all of those who have already decided on a No vote this early on, I'd like to collect your feedback about what is lacking from the 1.0 release of
Jakarta Data. (Thank you @hantsy for
pointing out that you want to see Reactive/Flow included). We still have 6 months of spec development timeframe at this point, and if you have concerns or specific requirements, it might be possible to add what you believe is lacking to version 1.0. Over-generalized
statements that the specification lacks maturity are not particularly helpful and are somewhat confusing given that the established products (Spring Data and Micronaut) are expected to become implementations of Jakarta Data. It would be much more helpful to
know exactly what from Spring Data and Micronaut that the Jakarta Data spec hasn't already standardized that you would like to see included.
Scott Stark wrote:
This spec is largely a copy of the Spring Data repository effort, and this is not something we fundamentally agree with. We advocate a active record
pattern to our users in our products, so fundamentally we disagree with this spec as a general approach for any profile or platform. Given that these cannot have an optional specification, the data spec would have to be getting traction in other EE implementations
and generating user interest before we would consider this for a profile or platform.
Nathan Rauh wrote:
Involvement from RedHat to help shape the specification would certainly be welcome. Its current focus on unifying Spring/Micronaut/JNoSQL follows from the contributions of participants.
3.
Platform-dev Archive: I could not find any discussion in the archive. There may be discussion, but I could not find it. Anyone with better
search skills please do provide links.
|
edburns@xxxxxxxxxxxxx | office:
+1 954 727 1095
| Calendar Booking:
https://aka.ms/meetedburns
|
| Please don't feel obliged to read or reply to this e-mail outside
| of your normal working hours.
|
| Reply anonymously to this email:
https://purl.oclc.org/NET/edburns/contact