Vetting JEP concepts?

[Andy Boothe]

Ethan, will respond directly!

Andy Boothe

On Mon, Apr 15, 2024 at 6:00 PM Ethan McCue <ethan at mccue.dev> wrote:

> Out of curiosity, what general subjects were you thinking you would submit
> JEPs for?
>
> On Mon, Apr 15, 2024, 1:56 PM Andy Boothe <andy.boothe at gmail.com> wrote:
>
>> Jesper, thank you for the additional context! I'll review the guide
>> straightaway.
>>
>> Andy Boothe
>>
>> On Mon, Apr 15, 2024 at 12:51 PM Jesper Wilhelmsson <
>> jesper.wilhelmsson at oracle.com> wrote:
>>
>>> Hi Andy,
>>>
>>> No, this is not the list to discuss any ideas about changes in the JDK.
>>> This list is for OpenJDK Community questions. The mailing list info page
>>> David linked to has short summaries of what each list is supposed to be
>>> used for.
>>>
>>> If you intend to start contributing to OpenJDK the best place to start
>>> is the OpenJDK Developers' Guide: https://openjdk.org/guide/
>>>
>>> There's plenty of advice there for how to proceed if you have an idea
>>> for a change in the JDK. If you have trouble finding the right list for
>>> some change you want to discuss you can find a Code Owners list at the end
>>> of that guide.
>>>
>>> Best Regards,
>>> /Jesper
>>>
>>> On Apr 15, 2024, at 18:18, Andy Boothe <andy.boothe at gmail.com> wrote:
>>>
>>> David and Thomas,
>>>
>>> Thank you both for the timely and insightful responses! They are greatly
>>> appreciated! This has helped me understand the process in much more detail.
>>>
>>> > Once you find a problem, next step is to hop onto the mailing list
>>> most relevant to the problem, and see if others in the community feel like
>>> it is as big/frequent of a problem that you say it is.
>>>
>>> Great! Thank you for sharing the list!
>>>
>>> > Good -- Why is ABC done this way (again, check ABC docs before
>>> hand)/When I tried doing ABC, I experienced pain point XYZ
>>> > Bad -- ABC feels poorly designed/ABC is wrong/ABC shouldn't be done
>>> this way -- Understand, literal thousands of people have looked over even
>>> the darkest corners of the JDK.
>>>
>>> Absolutely right. I completely agree, but this is an important point
>>> that bears repeating. Given that Java is one of - if not the - biggest
>>> platforms in the world, I know the likelihood that I'm the first person to
>>> magically see a problem is about 0%! And it's important to pursue any pain
>>> points or insights I have through the lens of the entire community, knowing
>>> that many folks have been focusing on this stuff for much longer than it
>>> took me to type out any email. :)
>>>
>>> > But JEPs come with a lot of red tape and require a significant amount
>>> of investment from other parties. Therefore, we try to avoid JEPs if
>>> possible, opting instead for more low-key processes (e.g., RFEs with CSRs).
>>>
>>> Oof! I certainly didn't mean to start with the nuclear option! Clearly I
>>> need to re-read JEP 1 a little more closely. Thank you for pointing that
>>> out, Thomas!
>>>
>>> Given these answers, I have a couple more follow-up questions, then
>>> hopefully I can start being a little more independent:
>>>
>>> 1. To David's point, I (strongly) suspect that some/most/all of my ideas
>>> have been considered and discussed before on this mailing list. What is the
>>> best way to search the mail archives? I've done some Google searching and
>>> came across gmane.io and
>>> https://hixon10.github.io/openjdk-mailing-lists-search/. Are these the
>>> preferred method(s)?
>>> 2. Forgive me if this is a stupid question, but would this not a
>>> suitable mailing list to start discussion, at least for some ideas?
>>> Naturally, if my question is about Valhalla, then I would go to
>>> valhalla-dev, but what about for more general Java-level questions?
>>>
>>> Thank you again!
>>>
>>> Andy Boothe
>>>
>>> On Mon, Apr 15, 2024 at 2:30 AM Thomas Stüfe <thomas.stuefe at gmail.com>
>>> wrote:
>>>
>>>> Hi Andy,
>>>>
>>>> David wrote a very good answer, I just wanted to expand on the last
>>>> point a bit:
>>>>
>>>> JEPs are a means, not an end. One typically does not set out to write a
>>>> JEP. Instead, one usually starts with a problem one wants to solve and
>>>> discusses this on the mailing lists.
>>>>
>>>> If, during the discussion, we see that the problem fits certain
>>>> criteria, a JEP would be needed. But JEPs come with a lot of red tape and
>>>> require a significant amount of investment from other parties. Therefore,
>>>> we try to avoid JEPs if possible, opting instead for more low-key processes
>>>> (e.g., RFEs with CSRs).
>>>>
>>>> Cheers, Thomas
>>>>
>>>> On Sun, Apr 14, 2024 at 10:45 PM David Alayachew <
>>>> davidalayachew at gmail.com> wrote:
>>>>
>>>>> Hello Andy. Thanks for getting involved.
>>>>>
>>>>> As for how to approach, there's a few steps involved before writing up
>>>>> or introducing JEP drafts.
>>>>>
>>>>> First thing is to find a problem that is either big enough or frequent
>>>>> enough that a change makes sense. As for what a big/frequent enough problem
>>>>> looks like, look at some of the JEP's -- in-flight, or delivered, that
>>>>> address the problems.
>>>>>
>>>>> Once you find a problem, next step is to hop onto the mailing list
>>>>> most relevant to the problem, and see if others in the community feel like
>>>>> it is as big/frequent of a problem that you say it is. Here is the list of
>>>>> mailing lists -- https://mail.openjdk.org/mailman/listinfo
>>>>>
>>>>> And if you want some traction, maybe cross post to the Java subreddit
>>>>> (with mod approval beforehand!). As long as you are respectful, read the
>>>>> documentation so you come informed, and don't jump to conclusions, you
>>>>> should find helpful responses in both places. Examples below.
>>>>>
>>>>> Good -- Why is ABC done this way (again, check ABC docs before
>>>>> hand)/When I tried doing ABC, I experienced pain point XYZ
>>>>>
>>>>> Bad -- ABC feels poorly designed/ABC is wrong/ABC shouldn't be done
>>>>> this way -- Understand, literal thousands of people have looked over even
>>>>> the darkest corners of the JDK.
>>>>>
>>>>> If your problem finds traction, next step is bringing up ideas on how
>>>>> to solve it. This is not where you start typing up a JEP format. Instead,
>>>>> introduce a solution and see if it gets anywhere. Fair warning, this is a
>>>>> higher bar to reach, and is usually as far as most ideas go.
>>>>>
>>>>> If you manage to cross that bar, post again and someone here will be
>>>>> happy to show you next steps.
>>>>>
>>>>> Let me know if you have any questions.
>>>>>
>>>>> Thank you for reaching out!
>>>>> David Alayachew
>>>>>
>>>>>
>>>>> On Sun, Apr 14, 2024 at 12:47 PM Andy Boothe <andy.boothe at gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> I have a couple of ideas for JEPs. What is the best way to vet these
>>>>>> concepts before I submit them to the official process? I'd like to confirm
>>>>>> that these ideas haven't already been entertained before (I have already
>>>>>> checked the existing draft JEPs), and I'd also like to make sure they are
>>>>>> even interesting and viable before I waste anyone's time on the official
>>>>>> committees.
>>>>>>
>>>>>> Thank you in advance for all feedback!
>>>>>>
>>>>>> Andy Boothe
>>>>>> *Email*: andy.boothe at gmail.com
>>>>>> *Mobile*: (979) 574-1089
>>>>>>
>>>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/discuss/attachments/20240415/9c2bd597/attachment.htm>