You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Ole Ersoy <ol...@gmail.com> on 2014/10/17 20:49:29 UTC
[Math] Github vs. ASF Repo vs. Rest of World (Non Apache Github Users)
Hi,
I'm following the discussion of how MATH-1138 was handled (Which I enjoy reading because I'm very impressed with how eloquently everyone communicates their points of view).
Just a warning that I might be ignoring [2] (Stolen from Gilles), because I have suggested this before:
==================================
Keep discussions in Github issues
==================================
This would have the following benefits:
- All Github users can follow the project just by clicking "Watch" (Which is what a lot of developers (World outside of Apache) are becoming accustomed to, and they might be confused by the fact that this is not the workflow)
- The mailing list could watch the Github issue so that trivial discussions could be kept on the list, and the really important points / final decisions summarized in the issue.
- It makes it easy to lookup workflow history. All the information / history is bound to the issue (Design, debate, links to pull requests, etc.).
- It eliminates missed communication.
- The commons math communication would be automatically filtered (Currently it's bundled with all Commons communication)
- You can have it all:
https://github.com/joeyh/github-backup
Even though the agreed upon protocol right now is to discuss everything on the mailing list, in practice it's both the mailing list and issues, and this is causing the type of extra work effort we are seeing with respect to MATH-1138, in addition to every now and then individuals have to be reminded to move discussion from JIRA to the list.
[2] This is already after the self-censorship filter, on issues
where I know in advance that challenging the adopted view will
either be ignored or go nowhere... :-}
Cheers,
- Ole
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Math] Github vs. ASF Repo vs. Rest of World (Non Apache Github Users)
Posted by Hank Grabowski <ha...@applieddefense.com>.
Ah, that is a more complicated question. Those discussions definitely
aren't part of the repository. As someone now recently bitten by the
confusion in where discussions should be taking place, I agree there should
be a good documentation of those swim lanes. I'm new to contributing to
open source projects, so I unfortunately don't have any past experience to
how I would handle the information flow. I thought that the original flow
of discussions in mailing lists leading to JIRA incidents and then
capturing the development notes within the JIRA incident, including the
pull request discussions, were good. Obviously that didn't work out as
well in practice as I had hoped.
On Fri, Oct 17, 2014 at 4:24 PM, sebb <se...@gmail.com> wrote:
> On 17 October 2014 21:16, Hank Grabowski <ha...@applieddefense.com> wrote:
> > The nice thing about Github, from my perspective as a person with only
> > read-only access to the ASF repositories, is that it provides me with the
> > ability to work in my own fork and then initiate pull requests that can
> be
> > incorporated into the root repository. I think it is still ideal that
> the
> > GitHub repository be synchronized with the ASF git repository and that
> the
> > ASF git repository be considered the "gold" copy of all things Apache.
> The
> > good news is that because this is git and not SVN, all changes from
> across
> > all the users in both GitHub land and direct ASF contributors will always
> > exist in everyone's repository, because of the nature of how git keeps
> > track of changes. We essentially all have a full copy of the repository
> > that can be used as a master to clone from if it ever came to it.
> >
>
> I was referring to the use of Github for discussing issues.
> AFAIK, issues are not tracked in the repository.
> So if Github disappears, would it still be possible to reconstruct the
> discussions?
>
> >
> > On Fri, Oct 17, 2014 at 4:12 PM, sebb <se...@gmail.com> wrote:
> >
> >> On 17 October 2014 19:49, Ole Ersoy <ol...@gmail.com> wrote:
> >> > Hi,
> >> >
> >> > I'm following the discussion of how MATH-1138 was handled (Which I
> enjoy
> >> > reading because I'm very impressed with how eloquently everyone
> >> communicates
> >> > their points of view).
> >> >
> >> > Just a warning that I might be ignoring [2] (Stolen from Gilles),
> >> because I
> >> > have suggested this before:
> >> > ==================================
> >> > Keep discussions in Github issues
> >> > ==================================
> >> >
> >> > This would have the following benefits:
> >> > - All Github users can follow the project just by clicking "Watch"
> >> (Which is
> >> > what a lot of developers (World outside of Apache) are becoming
> >> accustomed
> >> > to, and they might be confused by the fact that this is not the
> workflow)
> >> > - The mailing list could watch the Github issue so that trivial
> >> discussions
> >> > could be kept on the list, and the really important points / final
> >> decisions
> >> > summarized in the issue.
> >> > - It makes it easy to lookup workflow history. All the information /
> >> > history is bound to the issue (Design, debate, links to pull requests,
> >> > etc.).
> >> > - It eliminates missed communication.
> >> > - The commons math communication would be automatically filtered
> >> (Currently
> >> > it's bundled with all Commons communication)
> >> > - You can have it all:
> >> > https://github.com/joeyh/github-backup
> >> >
> >> > Even though the agreed upon protocol right now is to discuss
> everything
> >> on
> >> > the mailing list, in practice it's both the mailing list and issues,
> and
> >> > this is causing the type of extra work effort we are seeing with
> respect
> >> to
> >> > MATH-1138, in addition to every now and then individuals have to be
> >> reminded
> >> > to move discussion from JIRA to the list.
> >>
> >> A major disadvantage is that Github is external.
> >> It may not last as long as the ASF.
> >> If it disappears or moves, all the history is potentially lost.
> >>
> >> > [2] This is already after the self-censorship filter, on issues
> >> > where I know in advance that challenging the adopted view will
> >> > either be ignored or go nowhere... :-}
> >> >
> >> > Cheers,
> >> > - Ole
> >> >
> >> > ---------------------------------------------------------------------
> >> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> >> > For additional commands, e-mail: dev-help@commons.apache.org
> >> >
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> >> For additional commands, e-mail: dev-help@commons.apache.org
> >>
> >>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
Re: [Math] Github vs. ASF Repo vs. Rest of World (Non Apache Github Users)
Posted by sebb <se...@gmail.com>.
On 17 October 2014 21:16, Hank Grabowski <ha...@applieddefense.com> wrote:
> The nice thing about Github, from my perspective as a person with only
> read-only access to the ASF repositories, is that it provides me with the
> ability to work in my own fork and then initiate pull requests that can be
> incorporated into the root repository. I think it is still ideal that the
> GitHub repository be synchronized with the ASF git repository and that the
> ASF git repository be considered the "gold" copy of all things Apache. The
> good news is that because this is git and not SVN, all changes from across
> all the users in both GitHub land and direct ASF contributors will always
> exist in everyone's repository, because of the nature of how git keeps
> track of changes. We essentially all have a full copy of the repository
> that can be used as a master to clone from if it ever came to it.
>
I was referring to the use of Github for discussing issues.
AFAIK, issues are not tracked in the repository.
So if Github disappears, would it still be possible to reconstruct the
discussions?
>
> On Fri, Oct 17, 2014 at 4:12 PM, sebb <se...@gmail.com> wrote:
>
>> On 17 October 2014 19:49, Ole Ersoy <ol...@gmail.com> wrote:
>> > Hi,
>> >
>> > I'm following the discussion of how MATH-1138 was handled (Which I enjoy
>> > reading because I'm very impressed with how eloquently everyone
>> communicates
>> > their points of view).
>> >
>> > Just a warning that I might be ignoring [2] (Stolen from Gilles),
>> because I
>> > have suggested this before:
>> > ==================================
>> > Keep discussions in Github issues
>> > ==================================
>> >
>> > This would have the following benefits:
>> > - All Github users can follow the project just by clicking "Watch"
>> (Which is
>> > what a lot of developers (World outside of Apache) are becoming
>> accustomed
>> > to, and they might be confused by the fact that this is not the workflow)
>> > - The mailing list could watch the Github issue so that trivial
>> discussions
>> > could be kept on the list, and the really important points / final
>> decisions
>> > summarized in the issue.
>> > - It makes it easy to lookup workflow history. All the information /
>> > history is bound to the issue (Design, debate, links to pull requests,
>> > etc.).
>> > - It eliminates missed communication.
>> > - The commons math communication would be automatically filtered
>> (Currently
>> > it's bundled with all Commons communication)
>> > - You can have it all:
>> > https://github.com/joeyh/github-backup
>> >
>> > Even though the agreed upon protocol right now is to discuss everything
>> on
>> > the mailing list, in practice it's both the mailing list and issues, and
>> > this is causing the type of extra work effort we are seeing with respect
>> to
>> > MATH-1138, in addition to every now and then individuals have to be
>> reminded
>> > to move discussion from JIRA to the list.
>>
>> A major disadvantage is that Github is external.
>> It may not last as long as the ASF.
>> If it disappears or moves, all the history is potentially lost.
>>
>> > [2] This is already after the self-censorship filter, on issues
>> > where I know in advance that challenging the adopted view will
>> > either be ignored or go nowhere... :-}
>> >
>> > Cheers,
>> > - Ole
>> >
>> > ---------------------------------------------------------------------
>> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> > For additional commands, e-mail: dev-help@commons.apache.org
>> >
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> For additional commands, e-mail: dev-help@commons.apache.org
>>
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Math] Github vs. ASF Repo vs. Rest of World (Non Apache Github Users)
Posted by Hank Grabowski <ha...@applieddefense.com>.
The nice thing about Github, from my perspective as a person with only
read-only access to the ASF repositories, is that it provides me with the
ability to work in my own fork and then initiate pull requests that can be
incorporated into the root repository. I think it is still ideal that the
GitHub repository be synchronized with the ASF git repository and that the
ASF git repository be considered the "gold" copy of all things Apache. The
good news is that because this is git and not SVN, all changes from across
all the users in both GitHub land and direct ASF contributors will always
exist in everyone's repository, because of the nature of how git keeps
track of changes. We essentially all have a full copy of the repository
that can be used as a master to clone from if it ever came to it.
On Fri, Oct 17, 2014 at 4:12 PM, sebb <se...@gmail.com> wrote:
> On 17 October 2014 19:49, Ole Ersoy <ol...@gmail.com> wrote:
> > Hi,
> >
> > I'm following the discussion of how MATH-1138 was handled (Which I enjoy
> > reading because I'm very impressed with how eloquently everyone
> communicates
> > their points of view).
> >
> > Just a warning that I might be ignoring [2] (Stolen from Gilles),
> because I
> > have suggested this before:
> > ==================================
> > Keep discussions in Github issues
> > ==================================
> >
> > This would have the following benefits:
> > - All Github users can follow the project just by clicking "Watch"
> (Which is
> > what a lot of developers (World outside of Apache) are becoming
> accustomed
> > to, and they might be confused by the fact that this is not the workflow)
> > - The mailing list could watch the Github issue so that trivial
> discussions
> > could be kept on the list, and the really important points / final
> decisions
> > summarized in the issue.
> > - It makes it easy to lookup workflow history. All the information /
> > history is bound to the issue (Design, debate, links to pull requests,
> > etc.).
> > - It eliminates missed communication.
> > - The commons math communication would be automatically filtered
> (Currently
> > it's bundled with all Commons communication)
> > - You can have it all:
> > https://github.com/joeyh/github-backup
> >
> > Even though the agreed upon protocol right now is to discuss everything
> on
> > the mailing list, in practice it's both the mailing list and issues, and
> > this is causing the type of extra work effort we are seeing with respect
> to
> > MATH-1138, in addition to every now and then individuals have to be
> reminded
> > to move discussion from JIRA to the list.
>
> A major disadvantage is that Github is external.
> It may not last as long as the ASF.
> If it disappears or moves, all the history is potentially lost.
>
> > [2] This is already after the self-censorship filter, on issues
> > where I know in advance that challenging the adopted view will
> > either be ignored or go nowhere... :-}
> >
> > Cheers,
> > - Ole
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> > For additional commands, e-mail: dev-help@commons.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
Re: [Math] Github vs. ASF Repo vs. Rest of World (Non Apache Github
Users)
Posted by Ole Ersoy <ol...@gmail.com>.
On 10/17/2014 03:12 PM, sebb wrote:
> On 17 October 2014 19:49, Ole Ersoy <ol...@gmail.com> wrote:
>> Hi,
>>
>> I'm following the discussion of how MATH-1138 was handled (Which I enjoy
>> reading because I'm very impressed with how eloquently everyone communicates
>> their points of view).
>>
>> Just a warning that I might be ignoring [2] (Stolen from Gilles), because I
>> have suggested this before:
>> ==================================
>> Keep discussions in Github issues
>> ==================================
>>
>> This would have the following benefits:
>> - All Github users can follow the project just by clicking "Watch" (Which is
>> what a lot of developers (World outside of Apache) are becoming accustomed
>> to, and they might be confused by the fact that this is not the workflow)
>> - The mailing list could watch the Github issue so that trivial discussions
>> could be kept on the list, and the really important points / final decisions
>> summarized in the issue.
>> - It makes it easy to lookup workflow history. All the information /
>> history is bound to the issue (Design, debate, links to pull requests,
>> etc.).
>> - It eliminates missed communication.
>> - The commons math communication would be automatically filtered (Currently
>> it's bundled with all Commons communication)
>> - You can have it all:
>> https://github.com/joeyh/github-backup
>>
>> Even though the agreed upon protocol right now is to discuss everything on
>> the mailing list, in practice it's both the mailing list and issues, and
>> this is causing the type of extra work effort we are seeing with respect to
>> MATH-1138, in addition to every now and then individuals have to be reminded
>> to move discussion from JIRA to the list.
> A major disadvantage is that Github is external.
> It may not last as long as the ASF.
> If it disappears or moves, all the history is potentially lost.
All of us and the ASF can backup of everything:
https://github.com/joeyh/github-backup
It also looks like (Need to play with it) commit hooks can be used to sync the apache git repository, ensure that the apache mailing list is update with all communication, and trigger backups to multiple locations:
http://git-scm.com/book/en/Customizing-Git-Git-Hooks
Some projects hooks to trigger instant web site updates whenever the documentation is updated.
Cheers,
- Ole
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Math] Github vs. ASF Repo vs. Rest of World (Non Apache Github Users)
Posted by sebb <se...@gmail.com>.
On 17 October 2014 19:49, Ole Ersoy <ol...@gmail.com> wrote:
> Hi,
>
> I'm following the discussion of how MATH-1138 was handled (Which I enjoy
> reading because I'm very impressed with how eloquently everyone communicates
> their points of view).
>
> Just a warning that I might be ignoring [2] (Stolen from Gilles), because I
> have suggested this before:
> ==================================
> Keep discussions in Github issues
> ==================================
>
> This would have the following benefits:
> - All Github users can follow the project just by clicking "Watch" (Which is
> what a lot of developers (World outside of Apache) are becoming accustomed
> to, and they might be confused by the fact that this is not the workflow)
> - The mailing list could watch the Github issue so that trivial discussions
> could be kept on the list, and the really important points / final decisions
> summarized in the issue.
> - It makes it easy to lookup workflow history. All the information /
> history is bound to the issue (Design, debate, links to pull requests,
> etc.).
> - It eliminates missed communication.
> - The commons math communication would be automatically filtered (Currently
> it's bundled with all Commons communication)
> - You can have it all:
> https://github.com/joeyh/github-backup
>
> Even though the agreed upon protocol right now is to discuss everything on
> the mailing list, in practice it's both the mailing list and issues, and
> this is causing the type of extra work effort we are seeing with respect to
> MATH-1138, in addition to every now and then individuals have to be reminded
> to move discussion from JIRA to the list.
A major disadvantage is that Github is external.
It may not last as long as the ASF.
If it disappears or moves, all the history is potentially lost.
> [2] This is already after the self-censorship filter, on issues
> where I know in advance that challenging the adopted view will
> either be ignored or go nowhere... :-}
>
> Cheers,
> - Ole
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org