Reduce the Jakarta NoSQL only for High level API

[otaviojava]

This PR reduces the Jakarta NoSQL scope for high-level abstraction. It will make it easier to have more implementation.

To Eclipse JNoSQL users, it won't impact anything once it goes to JNoSQL implementation.

Changes

• Add Jakarta Data as API
• Remove communication as API
• Remove Repository to use Jakarta Data instead

⚠️ for JNoSQL users, it will impact the package name. "jakarta.nosql"

Ref: #153

[keilw]

Why was the DatabaseType removed?
Is there another way to tell which DB type you're using?

[keilw]

Crippling and disabling much of the API without a proper replacement is not a good approach.
Why wasn't most of it parked e.g. in the https://github.com/JNOSQL orga instead of deleting it all?

Beside a significant API reduction must go via a Release Review by the Spec Committee first.

[otaviojava]

@keilw those are a good point, I'll move it to JNoQL and show the full show later next week.

[keilw]

@otaviojava Please wait till approached by @ivargrimstad or another mentor/Spec Committee Member.
Removing such big chunks of the API/SPI should not be done without a review.
And I saw, that except for a progress review and ballot last October, there have not been any release reviews for either Jakarta NoSQL or JNoSQL.

While it may not affect end users/developers using the API removing all those "SPIs" and stuffing them into the only compatible implementation could have a negative impact on other possible implementors, which is why those members should get a chance to review the proposed changes first.

[otaviojava]

@keilw

Beside a significant API reduction must go via a Release Review by the Spec Committee first.

Yes that is the next step.

Ps: the release review happened through a milestone release, not a PR. And we have one year to give progress:

Progress reviews are not part of the normal flow of the JESP: creation->plan->release->plan->release->plan->release and so on.
A progress review can be initiated by the project when they want to inform the specification committee about their progress, but are not ready for a release yet.
A progress review may be requested by the specification committee if no progress has been made in 12 months.
Read more about progress reviews in the EFSP.

Reference:
https://jakarta.ee/committees/specification/guide/

[keilw]

@otaviojava All of these were always done without a review, JNoSQL had the last release review on May 15th, 2019: https://projects.eclipse.org/projects/technology.jnosql
Anything that gets published to https://central.sonatype.dev/search?q=jakarta.nosql and not just a snapshot repo should have a review.

A progress review may be requested by the specification committee if no progress has been made in 12 months.

That means if you did not do one in one year it may be requested by the spec committee, but you're supposed to request a review from the project side, if you publish a milestone, especially with such a massive change.
Jakarta Config just started with a totally clean slate, and maybe even that's something to explore for NoSQL instead of cutting half of it away ;-)

Either way, Config scheduled what seems like a progress review in May, so we should probably do the same here.

[ivargrimstad]

A couple of comments:

• I think the changes are within the scope of the spec project, so you are all good there

• A progress review about a year after the previous one is sufficient (unless you are close to a 1.0 release review then)

• JNoSQL is a separate open source project under the Technology PMC and does not need to be mixed into this discussion. Please refer to that PMC with anything regarding release reviews of such of that project as it is not a concern of the Jakarta EE Specification Committee or the EE4J PMC

[otaviojava]

Furthermore, @keilw, take a look at here:

jakartaee/specifications#587 (comment)