We have a few decisions here.

1)  Jorn's sandbox idea would probably be a snapshot of the current
trunk.  At least that is what I think he is proposing.

2)  Your idea is also possible... with a branch, then we also have the
ability to work and keep two branches up to date if needed.  The 1.5.3
branch could be patched with critical fixes we develop in the trunk and
allow us to release more...  kinda like what Apache HTTP server did with
1.4.3 and the newer branches.  Only problems with it are (a) how long do
we maintain the branch and (b) it becomes a management workload to
determine what goes where and when to release.

I'm going to think about it for a bit.

James

On 5/4/2012 9:39 AM, [email protected] wrote:
> How would the sandbox release work? We would create a branch from the trunk?
>
> I can't be the release manager for 1.5.3, although I would enjoy the
> experience. I am up to my neck with work for my thesis.
>
> On Fri, May 4, 2012 at 8:20 AM, Jörn Kottmann <[email protected]> wrote:
>
>> We should make a sandbox release of it first.
>> In my opinion that has a few advantages:
>> - we can get user feedback
>> - it can be released more often and is not bound to the longer opennlp
>> cycle
>> - its easier to change it (e.g. breaking APi changes)
>>
>> Jörn
>>
>>
>> On 05/03/2012 06:16 AM, Boris Galitsky wrote:
>>
>>> Hi James
>>> Should Similarity component come with 1.5.3?
>>> It has been stable for more than a month now and needs just a final touch
>>> RegardsBoris
>>>
>>>
>>>
>>>
>>>
>>>  Date: Wed, 2 May 2012 22:50:12 -0400
>>>> From: [email protected]
>>>> To: [email protected]
>>>> Subject: OpenNLP 1.5.3 ....
>>>>
>>>> Everyone,
>>>>
>>>> Does anyone object to us releasing a 1.5.3 release?
>>>> I'm only asking because we have several important fixes currently in
>>>> trunk that some people are coming across as they are trying to use and
>>>> we also haven't had a release since graduating.
>>>> I'm not trying to push the issue; but, would be nice.
>>>>
>>>> I've also got more data we can test against with the addition of the
>>>> ConLL 02 and ConLLX data I downloaded that is available.
>>>>
>>>> Thanks,
>>>> James
>>>>
>>>
>>

Reply via email to