Hi all,

as a follow-up I’ve created OPENNLP-1708 and a path for OpenNLP 3.x, see: 
https://issues.apache.org/jira/browse/OPENNLP-1708

Jira login required for all details.

Feel free to add additional hints/pointers, create new sub-tasks and share 
creative ideas.

Gruß
Martin
—

> Am 12.03.2025 um 18:31 schrieb Martin Wiesner <mawie...@apache.org>:
> 
> Thanks Jeff for sharing your thoughts on this. 
> I think, we should aim for backward compatibility as much as possible.
> 
> Since no other feedback arrived (so far) for this thread, we’ll assume lazy 
> consensus for the outlined modularization of Apache OpenNLP for the target 
> release 3.x and onwards.
> 
> A new (epic) Jira issue and related branch will therefore be opened soon.
> I’ll inform the list again, once both things have happened.
> 
> Gruß
> Martin
> --
> 
>> Am 14.02.2025 um 22:26 schrieb Jeff Zemerick <jzemer...@apache.org>:
>> 
>> I think that's a good path to go down.
>> 
>> I also think it would be ok to *not* maintain compatibility with the
>> existing CLI in places where things can be improved. While backward
>> compatibility is helpful, a new major release implies there may be breaking
>> changes. I don't have a good feel for how many users there are of the CLI
>> but if I had to guess I would expect it to be somewhat low in comparison to
>> users who integrate OpenNLP in their code. I'm happy to have my guess
>> challenged though.
>> 
>> Would this be a good time to take a look at the things in the
>> opennlp-sandbox and see if anything can be moved over to the main project?
>> 
>> Thanks,
>> Jeff
>> 
>> 
>> On Thu, Feb 13, 2025 at 7:17 AM Richard Zowalla <r...@apache.org> wrote:
>> 
>>> FYI: Have invited the user list last week:
>>> https://lists.apache.org/thread/wgj7oovl4ls7pmpbvbc9hoo895whjpd5
>>> 
>>> On 2025/02/13 11:35:49 Martin Wiesner wrote:
>>>> Hi all,
>>>> 
>>>> thanks Bruno for sharing your thoughts.
>>>> 
>>>> Obviously, I’m a +1 for the idea of a more modularized form of opennlp
>>> core components, see Richard's original post.
>>>> 
>>>> To get things rolling, I’d suggest:
>>>> The feedback and early conceptual phase could be open until say March
>>> 1st 2025. Thoughts?
>>>> 
>>>> Any more feedback is highly welcome. Should we, in addition, share the
>>> modularization proposal (3.x) on the users list so the community can join
>>> the discussion?
>>>> 
>>>> Please participate in this discussion, especially if you are a senior
>>> committer and/or PMC member of the OpenNLP project.
>>>> 
>>>> Gruß
>>>> Martin
>>>> —
>>>> 
>>>>> Am 09.02.2025 um 10:09 schrieb Bruno Kinoshita <
>>> brunodepau...@gmail.com>:
>>>>> 
>>>>> Sounds like a good plan, especially "We also aim to maintain backward
>>>>> compatibility" 👍
>>>>> +1
>>>>> 
>>>>> Thanks!
>>>>> 
>>>>> 
>>>>> On Sat, 8 Feb 2025 at 11:06, Richard Zowalla <r...@apache.org> wrote:
>>>>> 
>>>>>> Hi,
>>>>>> 
>>>>>> Martin and I were sitting on the ICE from Berlin back to Heilbronn
>>> and had
>>>>>> a discussion about modernizing OpenNLP's module structure.
>>>>>> We also discussed that idea with Atita off-list in Berlin two days
>>> ago :-)
>>>>>> 
>>>>>> Our main goal is to split the CLI parts and define a solid API module,
>>>>>> along with well-structured dependent modules something like „api“,
>>> „cli“,
>>>>>> „core“, „addons“, ...
>>>>>> 
>>>>>> We also aim to maintain backward compatibility for the CLI, ensuring
>>> that
>>>>>> users can still grab the distribution binary or „opennlp-tools“ via
>>> Maven
>>>>>> as before.
>>>>>> 
>>>>>> By implementing these changes, we believe OpenNLP will become more
>>>>>> modular, flexible, and easier to maintain.
>>>>>> This should make it easier to integrate OpenNLP into core libraries
>>> while
>>>>>> improving maintainability for future use cases.
>>>>>> 
>>>>>> Our proposal would be, that we start with OpenNLP 3.x on a separate
>>> branch
>>>>>> to try things out and call this a step towards the 3.x series of
>>> OpenNLP.
>>>>>> 
>>>>>> WDYT?
>>>>>> 
>>>>>> Gruß
>>>>>> Richard
>>>> 
>>>> 
>>> 
> 

Reply via email to