You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@datasketches.apache.org by Evans Ye <ev...@apache.org> on 2019/10/23 08:57:16 UTC

The way to communicate with datasketches users

Hi team,

I was exploring the functionality of DataSketches and I got a question I
want to ask. I found right now we use Slack and Google Forum for user
interaction:

* Google: https://groups.google.com/forum/#!forum/sketches-user
* Slack: https://app.slack.com/client/T4S1WH2J3/CHWKFEHUZ

I wonder what's the preferred way for the community?
I think Slack might not be a good way for "StackOverflow" like Q&A, nor
good for searching and cross timezone collaboration, but is very efficient
for deep-dive discussions. Google is OK but it does not integrate with
ASF's tool for stats and archiving.

So, what I'm thinking is to create a user@datasketches.apache.org mailing
list for users. What do the community think?

Best,
Evans Ye

Re: The way to communicate with datasketches users

Posted by Furkan KAMACI <fu...@gmail.com>.
Hi Lee,

I understand your concerns. As I mentioned earlier, our mottos at Apache
are:

"If it didn't happen on a mailing list, it didn't happen."
"Put community before code."

So, the mail list is important. However, documentation, development
guidelines and FAQs are important too for a project to help its community.
I think that the problems of getting rid of Google Forum can be handled.

On the other hand, I see some other ASF projects too which want to
communicate via Slack or alternatives. I understand that need too, as an
example, Pulsar sends daily digests into their subscribers. It (may)
increase the speed of communication, but searchability gets hardened and
messages become bad structured.

Kind Regards,
Furkan KAMACI

On Mon, Oct 28, 2019 at 8:38 PM Evans Ye <ev...@apache.org> wrote:

> Thanks for the summary, Lee.
>
> To clarify my intention, I propose the user mailing list out of the goal
> of making DataSketches graduated to TLP.
> Apache mailing list might has drawbacks comparing to the others. But it's
> a consistent way for user discussions across all ASF projects.
> Moreover, I have an argument from the ASF's point of view:
>
> To be an TLP operating under ASF, it's important to show the user that the
> project has enough diversity so that no one single player can greatly
> impact the project. On the other side, the project should also be
> sustainable under ASF's support w/o being effected by other 3rd party orgs.
> If we're leverage other tools such as google forum as our official user
> discussion mailing list, what if google changed the policy or start to
> charge us? What if google decided not to continue to run forum? Can we get
> the full dump and import all the historical data back to ASF mailing list?
>
> Although the chance of my concern is tiny. I think my point stands.
>
> I'm neutral to your proposal of using Slack + archive daily digest  to our
> mailing list. ;)
> Comments are welcome.
>
> Best,
> Evans
>
> leerho <le...@gmail.com> 於 2019年10月24日 週四 上午4:53寫道:
>
>> Here is a summary of our Public Communications Channels for DataSketches:
>>
>> *Apache Mailing Lists at @datasketches.apache.org
>> <http://datasketches.apache.org>*
>>
>>    - commits
>>    - dev
>>    - issues              //for all users
>>    - notifications
>>
>> *the-asf.slack.com <http://the-asf.slack.com> channels*
>>
>>    - datasketches        //for all users
>>    - datasketches-dev
>>
>> *Google Forum for all users*
>>
>>    - https://groups.google.com/forum/#!topic/sketches-user
>>
>> *Every GitHub repository has an "issues" tab available to all users:*
>>
>>                                         Open    Closed
>> incubator-datasketches-java             4       31
>> incubator-datasketches-memory           1       45
>> incubator-datasketches-cpp              0       7
>> incubator-datasketches-hive             1       4
>> incubator-datasketches-pig              1       4
>> incubator-datasketches-postgresql       0       0
>> incubator-datasketches-vector           0       0
>> incubator-datasketches-characterization 0       0
>> incubator-datasketches-website          0       0
>>
>> *In addition, we also track issues on some major external sites where
>> they have integrated our library such as:*
>>
>>    - Apache Druid
>>    - Postgresql.org
>>
>> I'm not sure we need another one, unless we eliminate some.  The
>> advantage of the Google forum is that it allows separation by topic and is
>> searchable by Google search and keeps issues forever. The disadvantage of
>> the Apache mail lists is that it has no organization structure other than
>> threads (which is quite clumsy).  The issue lists in GitHub offer even more
>> organizational and cross referencing tools and are markdown compatible.
>>
>> If we were to have one primary user Q&A "channel" I would favor our new
>> slack channel "datasketches".  Especially once we figure out how to archive
>> daily digests.
>>
>> Comments?
>>
>> Lee.
>>
>>
>>
>> On Wed, Oct 23, 2019 at 11:52 AM Jon Malkin <jo...@gmail.com> wrote:
>>
>>> I certainly appreciate the idea, but I worry that in practice it will be
>>> just another way to reach us and need to be supported along with all the
>>> others. I am wary of adding yet another place that the team needs to keep
>>> monitoring.
>>>
>>> Lee has expressed interest in having a daily digest of the slack channel
>>> pushed to the mailing list, so maybe that would allow the channel contents
>>> to be treated as a searchable archive?
>>>
>>>   jon
>>>
>>> On Wed, Oct 23, 2019 at 1:57 AM Evans Ye <ev...@apache.org> wrote:
>>>
>>>> Hi team,
>>>>
>>>> I was exploring the functionality of DataSketches and I got a question
>>>> I want to ask. I found right now we use Slack and Google Forum for user
>>>> interaction:
>>>>
>>>> * Google: https://groups.google.com/forum/#!forum/sketches-user
>>>> * Slack: https://app.slack.com/client/T4S1WH2J3/CHWKFEHUZ
>>>>
>>>> I wonder what's the preferred way for the community?
>>>> I think Slack might not be a good way for "StackOverflow" like Q&A, nor
>>>> good for searching and cross timezone collaboration, but is very efficient
>>>> for deep-dive discussions. Google is OK but it does not integrate with
>>>> ASF's tool for stats and archiving.
>>>>
>>>> So, what I'm thinking is to create a user@datasketches.apache.org
>>>> mailing list for users. What do the community think?
>>>>
>>>> Best,
>>>> Evans Ye
>>>>
>>>

Re: The way to communicate with datasketches users

Posted by Evans Ye <ev...@apache.org>.
Thanks for the summary, Lee.

To clarify my intention, I propose the user mailing list out of the goal of
making DataSketches graduated to TLP.
Apache mailing list might has drawbacks comparing to the others. But it's a
consistent way for user discussions across all ASF projects.
Moreover, I have an argument from the ASF's point of view:

To be an TLP operating under ASF, it's important to show the user that the
project has enough diversity so that no one single player can greatly
impact the project. On the other side, the project should also be
sustainable under ASF's support w/o being effected by other 3rd party orgs.
If we're leverage other tools such as google forum as our official user
discussion mailing list, what if google changed the policy or start to
charge us? What if google decided not to continue to run forum? Can we get
the full dump and import all the historical data back to ASF mailing list?

Although the chance of my concern is tiny. I think my point stands.

I'm neutral to your proposal of using Slack + archive daily digest  to our
mailing list. ;)
Comments are welcome.

Best,
Evans

leerho <le...@gmail.com> 於 2019年10月24日 週四 上午4:53寫道:

> Here is a summary of our Public Communications Channels for DataSketches:
>
> *Apache Mailing Lists at @datasketches.apache.org
> <http://datasketches.apache.org>*
>
>    - commits
>    - dev
>    - issues              //for all users
>    - notifications
>
> *the-asf.slack.com <http://the-asf.slack.com> channels*
>
>    - datasketches        //for all users
>    - datasketches-dev
>
> *Google Forum for all users*
>
>    - https://groups.google.com/forum/#!topic/sketches-user
>
> *Every GitHub repository has an "issues" tab available to all users:*
>
>                                         Open    Closed
> incubator-datasketches-java             4       31
> incubator-datasketches-memory           1       45
> incubator-datasketches-cpp              0       7
> incubator-datasketches-hive             1       4
> incubator-datasketches-pig              1       4
> incubator-datasketches-postgresql       0       0
> incubator-datasketches-vector           0       0
> incubator-datasketches-characterization 0       0
> incubator-datasketches-website          0       0
>
> *In addition, we also track issues on some major external sites where they
> have integrated our library such as:*
>
>    - Apache Druid
>    - Postgresql.org
>
> I'm not sure we need another one, unless we eliminate some.  The advantage
> of the Google forum is that it allows separation by topic and is searchable
> by Google search and keeps issues forever. The disadvantage of the Apache
> mail lists is that it has no organization structure other than threads
> (which is quite clumsy).  The issue lists in GitHub offer even more
> organizational and cross referencing tools and are markdown compatible.
>
> If we were to have one primary user Q&A "channel" I would favor our new
> slack channel "datasketches".  Especially once we figure out how to archive
> daily digests.
>
> Comments?
>
> Lee.
>
>
>
> On Wed, Oct 23, 2019 at 11:52 AM Jon Malkin <jo...@gmail.com> wrote:
>
>> I certainly appreciate the idea, but I worry that in practice it will be
>> just another way to reach us and need to be supported along with all the
>> others. I am wary of adding yet another place that the team needs to keep
>> monitoring.
>>
>> Lee has expressed interest in having a daily digest of the slack channel
>> pushed to the mailing list, so maybe that would allow the channel contents
>> to be treated as a searchable archive?
>>
>>   jon
>>
>> On Wed, Oct 23, 2019 at 1:57 AM Evans Ye <ev...@apache.org> wrote:
>>
>>> Hi team,
>>>
>>> I was exploring the functionality of DataSketches and I got a question I
>>> want to ask. I found right now we use Slack and Google Forum for user
>>> interaction:
>>>
>>> * Google: https://groups.google.com/forum/#!forum/sketches-user
>>> * Slack: https://app.slack.com/client/T4S1WH2J3/CHWKFEHUZ
>>>
>>> I wonder what's the preferred way for the community?
>>> I think Slack might not be a good way for "StackOverflow" like Q&A, nor
>>> good for searching and cross timezone collaboration, but is very efficient
>>> for deep-dive discussions. Google is OK but it does not integrate with
>>> ASF's tool for stats and archiving.
>>>
>>> So, what I'm thinking is to create a user@datasketches.apache.org
>>> mailing list for users. What do the community think?
>>>
>>> Best,
>>> Evans Ye
>>>
>>

Re: The way to communicate with datasketches users

Posted by leerho <le...@gmail.com>.
Here is a summary of our Public Communications Channels for DataSketches:

*Apache Mailing Lists at @datasketches.apache.org
<http://datasketches.apache.org>*

   - commits
   - dev
   - issues              //for all users
   - notifications

*the-asf.slack.com <http://the-asf.slack.com> channels*

   - datasketches        //for all users
   - datasketches-dev

*Google Forum for all users*

   - https://groups.google.com/forum/#!topic/sketches-user

*Every GitHub repository has an "issues" tab available to all users:*

                                        Open    Closed
incubator-datasketches-java             4       31
incubator-datasketches-memory           1       45
incubator-datasketches-cpp              0       7
incubator-datasketches-hive             1       4
incubator-datasketches-pig              1       4
incubator-datasketches-postgresql       0       0
incubator-datasketches-vector           0       0
incubator-datasketches-characterization 0       0
incubator-datasketches-website          0       0

*In addition, we also track issues on some major external sites where they
have integrated our library such as:*

   - Apache Druid
   - Postgresql.org

I'm not sure we need another one, unless we eliminate some.  The advantage
of the Google forum is that it allows separation by topic and is searchable
by Google search and keeps issues forever. The disadvantage of the Apache
mail lists is that it has no organization structure other than threads
(which is quite clumsy).  The issue lists in GitHub offer even more
organizational and cross referencing tools and are markdown compatible.

If we were to have one primary user Q&A "channel" I would favor our new
slack channel "datasketches".  Especially once we figure out how to archive
daily digests.

Comments?

Lee.



On Wed, Oct 23, 2019 at 11:52 AM Jon Malkin <jo...@gmail.com> wrote:

> I certainly appreciate the idea, but I worry that in practice it will be
> just another way to reach us and need to be supported along with all the
> others. I am wary of adding yet another place that the team needs to keep
> monitoring.
>
> Lee has expressed interest in having a daily digest of the slack channel
> pushed to the mailing list, so maybe that would allow the channel contents
> to be treated as a searchable archive?
>
>   jon
>
> On Wed, Oct 23, 2019 at 1:57 AM Evans Ye <ev...@apache.org> wrote:
>
>> Hi team,
>>
>> I was exploring the functionality of DataSketches and I got a question I
>> want to ask. I found right now we use Slack and Google Forum for user
>> interaction:
>>
>> * Google: https://groups.google.com/forum/#!forum/sketches-user
>> * Slack: https://app.slack.com/client/T4S1WH2J3/CHWKFEHUZ
>>
>> I wonder what's the preferred way for the community?
>> I think Slack might not be a good way for "StackOverflow" like Q&A, nor
>> good for searching and cross timezone collaboration, but is very efficient
>> for deep-dive discussions. Google is OK but it does not integrate with
>> ASF's tool for stats and archiving.
>>
>> So, what I'm thinking is to create a user@datasketches.apache.org
>> mailing list for users. What do the community think?
>>
>> Best,
>> Evans Ye
>>
>

Re: The way to communicate with datasketches users

Posted by Jon Malkin <jo...@gmail.com>.
I certainly appreciate the idea, but I worry that in practice it will be
just another way to reach us and need to be supported along with all the
others. I am wary of adding yet another place that the team needs to keep
monitoring.

Lee has expressed interest in having a daily digest of the slack channel
pushed to the mailing list, so maybe that would allow the channel contents
to be treated as a searchable archive?

  jon

On Wed, Oct 23, 2019 at 1:57 AM Evans Ye <ev...@apache.org> wrote:

> Hi team,
>
> I was exploring the functionality of DataSketches and I got a question I
> want to ask. I found right now we use Slack and Google Forum for user
> interaction:
>
> * Google: https://groups.google.com/forum/#!forum/sketches-user
> * Slack: https://app.slack.com/client/T4S1WH2J3/CHWKFEHUZ
>
> I wonder what's the preferred way for the community?
> I think Slack might not be a good way for "StackOverflow" like Q&A, nor
> good for searching and cross timezone collaboration, but is very efficient
> for deep-dive discussions. Google is OK but it does not integrate with
> ASF's tool for stats and archiving.
>
> So, what I'm thinking is to create a user@datasketches.apache.org mailing
> list for users. What do the community think?
>
> Best,
> Evans Ye
>