Absolutely. We previously proposed the following, which is along the same
lines of separating the core SDK from add-on libraries.

* sdks/
  * java/
    * core/
    * io/
      * gcp/
      * kafka/
      * any_other_individual_io_library/
    * other_libraries_organized_by_groups/
  * py/
    * same_as_above

On Sat, Mar 5, 2016 at 11:29 PM, Jean-Baptiste Onofré <[email protected]>
wrote:

> By the way, related to BEAM-77, what do you think about creating a module
> for IO ?
> In addition of the existing IO (PubSub, Text, XML, ...), we plan to add
> new IOs (kafka, JMS, MQTT, ...). Instead to put directly in sdk/src, it
> would make sense to have:
>
> io module (pom)
> io/kafka
> io/jms
> ...
>
> It will give more visibility, and probably easier for contributions (as
> people can submit PR with a new IO as a module).
>
> Thoughts ?
>
> Regards
> JB
>
> On 03/05/2016 10:02 PM, Davor Bonaci wrote:
>
>> Already tracked by BEAM-78 [1].
>>
>> As per details there, we planned to postpone this rename just a little bit
>> -- until other code drops have been integrated into the repository, and we
>> have completed the refactoring that will separate the two uses of term
>> Dataflow -- one which needs to change and one that remains. Proceeding
>> right now might make this a little bit harder.
>>
>> Is this something we are comfortable with?
>>
>> Also, related issue is BEAM-77 which calls for directory reorganization.
>> Max was kind enough to do this from the start for the Flink runner -- we
>> need to do the same for the core SDK pieces.
>>
>> [1] https://issues.apache.org/jira/browse/BEAM-78
>>
>> On Sat, Mar 5, 2016 at 12:17 AM, Jean-Baptiste Onofré <[email protected]>
>> wrote:
>>
>> Hi guys,
>>>
>>> FYI, I'm preparing a PR to cleanup the POMs and reference Beam naming.
>>>
>>> For instance:
>>> - we should use Apache parent POM (it will simplify the release process)
>>> - the name in the pom should mention Beam instead of Google Cloud SDK
>>> - the groupId/artifactId should be org.apache.beam instead of
>>> com.google.cloud.dataflow
>>> - the developer should be the actual developer set (not google)
>>> - the scm should be the Apache git one (not Dataflow on github)
>>> - the legal headers should be Apache 2.0/ASF one
>>> - etc ...
>>>
>>> I'm preparing the PR that I will submit to you for review before pushing.
>>>
>>> Thanks !
>>> Regards
>>> JB
>>> --
>>> Jean-Baptiste Onofré
>>> [email protected]
>>> http://blog.nanthrax.net
>>> Talend - http://www.talend.com
>>>
>>>
>>
> --
> Jean-Baptiste Onofré
> [email protected]
> http://blog.nanthrax.net
> Talend - http://www.talend.com
>

Reply via email to