incubator-ooo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dave Fisher <dave2w...@comcast.net>
Subject Re: [DISCUSS] Pootle and New Contributor Category
Date Fri, 08 Jun 2012 01:29:41 GMT

On Jun 7, 2012, at 5:50 PM, Rob Weir wrote:

> On Thu, Jun 7, 2012 at 8:10 PM, Dave Fisher <dave2wave@comcast.net> wrote:
>> 
>> On Jun 7, 2012, at 4:30 PM, RGB ES wrote:
>> 
>>> 2012/6/8 Rob Weir <robweir@apache.org>:
>>>> On Thu, Jun 7, 2012 at 7:08 PM, Dave Fisher <dave2wave@comcast.net>
wrote:
>>>>> 
>>>>> On Jun 7, 2012, at 3:43 PM, Kay Schenk wrote:
>>>>> 
>>>>>> On Thu, Jun 7, 2012 at 10:24 AM, Dennis E. Hamilton <orcmid@apache.org>wrote:
>>>>>> 
>>>>>>> +1 on this discussion so far.
>>>>>>> 
>>>>>>> I was skeptical but I favor how this is going.
>>>>>>> 
>>>>>>> Also, the anonymous contribution to pootle is a no-no.
>>>>>>> 
>>>>>>> - Dennis
>>>>>>> 
>>>>>>> PS: Changing to the [DISCUSS] that is called for and to have
it be visible.
>>>>>>> 
>>>>>>> -----Original Message-----
>>>>>>> From: Rob Weir [mailto:robweir@apache.org]
>>>>>>> Sent: Thursday, June 07, 2012 09:41
>>>>>>> To: ooo-dev@incubator.apache.org
>>>>>>> Subject: Re: *DRAFT FINAL* June board report
>>>>>>> 
>>>>>>> On Thu, Jun 7, 2012 at 6:59 AM, Jürgen Schmidt
>>>>>>> <jogischmidt@googlemail.com> wrote:
>>>>>>>> On 6/7/12 12:10 PM, Ross Gardler wrote:
>>>>>>>>> On 7 June 2012 11:02, Jürgen Schmidt <jogischmidt@googlemail.com>
>>>>>>> wrote:
>>>>>>>>>> On 6/7/12 11:54 AM, Ross Gardler wrote:
>>>>>>>>>>> On 7 June 2012 10:47, Jürgen Schmidt <jogischmidt@googlemail.com>
>>>>>>> wrote:
>>>>>>>>>>>> On 6/7/12 11:28 AM, Ross Gardler wrote:
>>>>>>>>>>>>> On 7 June 2012 05:50, Herbert Duerr <hdu@apache.org>
wrote:
>>>>>>>>>>>>> 
>>>>>>>>>>>>> ...
>>>>>>>>>>>>> 
>>>>>>>>>>>>>> I think we maybe should add one more
topic here: Working with
>>>>>>> pootle
>>>>>>>>>>>>>> currently requires committership,
which results in translators
>>>>>>> having having
>>>>>>>>>>>>>> to be fast-tracked when they show
up on the mailing list. The
>>>>>>> board needs to
>>>>>>>>>>>>>> decide if this short-circuiting of
the process is desirable or not
>>>>>>> and what
>>>>>>>>>>>>>> the alternatives are.
>>>>>>>>>>>>> 
>>>>>>>>>>>>> No, need, that's not a board level issue.
It's up to the project to
>>>>>>> define its
>>>>>>>>>>>>> own expectations of committers.
>>>>>>>>>>>> 
>>>>>>>>>>>> it's a very bad limitation. I would prefer
a user management which
>>>>>>>>>>>> allows registration (by email verification)
of new users and where
>>>>>>> new
>>>>>>>>>>>> users agree to contribute under the Apache
license. Maybe combined
>>>>>>> with
>>>>>>>>>>>> an iCLA but not necessarily require to be
committer.
>>>>>>>>>>>> 
>>>>>>>>>>>> But I am not sure if something like that
would be possible at all.
>>>>>>>>>>>> 
>>>>>>>>>>>> Otherwise we have to deal with the current
approach and hope that we
>>>>>>> can
>>>>>>>>>>>> reach volunteers to accept this approach
and work together with them
>>>>>>> on
>>>>>>>>>>>> a fast-track.
>>>>>>>>>>> 
>>>>>>>>>>> I agree that the limitation suboptimal.
>>>>>>>>>>> 
>>>>>>>>>>> I suggest someone take this up with legal-discuss@
If legal@ feel
>>>>>>> able
>>>>>>>>>>> to approve a more relaxed approach to iCLAs for
access to Pootle then
>>>>>>>>>>> infra@ can be asked to find a technical solution.
>>>>>>>>>> 
>>>>>>>>>> I agree and thanks to remind me that I should take
the appropriate
>>>>>>>>>> action to address things like that ;-)
>>>>>>>>> 
>>>>>>>>> Careful with the "I" - madness lies that way ;-)
>>>>>>>>> 
>>>>>>>>> This is the perfect opportunity for someone lurking here
to make an
>>>>>>>>> early and potentially very significant contribution.
Shepherding these
>>>>>>>>> kinds of actions takes time away from those embedded
in the coding.
>>>>>>>>> It's a good way to earn merit while you figure out where
to contribute
>>>>>>>>> to the project. If someone like that is reading but not
sure how to
>>>>>>>>> proceed I'm sure others will help guide you.
>>>>>>>> 
>>>>>>>> I agree but the idea is not really new and nothing happened
so far ;-)
>>>>>>>> 
>>>>>>>> Thinking more about it I would like to discuss a new term
"Apache
>>>>>>>> contributor" where users can register for an user account
by accepting
>>>>>>>> that all their contributions are under ALv2. The verification
can be by
>>>>>>>> email verification and the iCLA can be required as well (details
have to
>>>>>>>> be defined). With such accounts people would get access to
more pubic
>>>>>>>> wikis (like our user wiki), tools like Pootle, bugzilla etc.
>>>>>>>> 
>>>>>>> 
>>>>>>> The "contributor" role at Apache already handles this.  A contributor
>>>>>>> can already register in Bugzilla, post patches, register in the
wiki,
>>>>>>> contribute documentation, etc.
>>>>>>> 
>>>>>>> What a contributor cannot do is directly modify the product code
in
>>>>>>> SVN.  So they are in RTC mode with respect to product code, including
>>>>>>> translations.
>>>>>>> 
>>>>>>> I think the disconnect here is we only have an anonymous method
for
>>>>>>> contributors to add translations to Pootle.  I can see the
>>>>>>> justification for requiring non-committers to submit translations
as
>>>>>>> patches in BZ or via suggestions in Pootle.  But the anonymous
part of
>>>>>>> this is completely wrong, both from community and from legal
>>>>>>> standpoint.
>>>>>>> 
>>>>>>> For example, those who contribute to Pootle, anonymously, see
their
>>>>>>> contributions marked as being from "nobody" in the UI:
>>>>>>> https://translate.apache.org/projects/OOo_34/
>>>>>>> 
>>>>>>> Isn't that rather insulting?
>>>>>>> 
>>>>>> 
>>>>>> [reposted since I didn't see this topic change]
>>>>>> 
>>>>>> yes, it is...I thought Juergen was suggesting that some special submission
>>>>>> access if you will be granted to the Pootle server.
>>>>> 
>>>>> As in we would like to be able to allow people with an iCLA on file to
register for access to the pootle server.
>>>>> 
>>>>> We can call these people "invited translators"
>>>>> 
>>>> 
>>>> Why not allow that to everyone?  I'm trying to see what harm would
>>>> come from that?  No one needs special permission to enter a BZ issue
>>>> and attach a patch.  Why can't someone log into Pootle and enter a
>>>> suggestion?  Is there a technical reason why this is not happening?
>>>> 
>>> 
>>> AFAIK, right now accept suggestions is not working but once this
>>> problem is solved I think that "opening" the system so anyone can
>>> issue a suggestion is really good: the analogy with BZ is just
>>> perfect.
>> 
>> Let's think of my idea as a possible compromise depending on what concerns Infrastructure
has with opening up pootle. Recall we are not the only project using this resource.
>> 
> 
> So long as your idea is to allowed logged in users to offer
> suggestions, then fine.  That is what RGB and I are say as well.

It is all about finding some way short of committer status to allow translators to make attributable
contributions. To go for less is trivializing the effort.

So, I'm not ignoring the rest of this thread. It is a good argument. I am looking for the
simplest transition. Raphael B and Jürgen have worked with Pootle perhaps they can say something
about what effort is needed to manage the pootle server from the PPMC side - the TC of the
RTC.

How much would user management add?

Regards,
Dave

> 
> But let's not reduce our view of the task of translation to such a
> level that we think this is a job that is trivial.   It is difficult,
> requires skill, and more skill than just knowing another language.  It
> can be done well, and it can be done poorly.  As such this skill
> should be demonstrated and observed within the community within a RTC
> environment before committership (free access to CRT) is offered.
> We should be as concerned with quality here as with any other part of
> the product.  As with code, some merit might be earned from effort on
> the legacy project.  But I don't think we solve any real problem by
> giving open access to CTR.
> 
> So I don't see this at all as being a Infra question.  And it has
> nothing to do with the iCLA.   It is more a RTC versus CTR question
> regarding core content of a release.  If someone has demonstrated
> merit, then we could make them a committer.  If they have not, then
> their contributions should be RTC.  I'm hoping that can be RTC with
> some sort of identity so we can track who they are and what their
> contributions have been.
> 
> Yes, finding a reviewer for some languages might be difficult.
> Finding good reviewers for critical sections of multi-threaded C++
> code is hard as well.  But we don't solve that problem by letting
> non-committers check in multi-threaded C++ code without a review
> because we think finding a reviewer is hard.
> 
> And remember, translation requires PO files, but it does not require
> Pootle.  Contributors who are new to the project can work off-line
> with the PO files, and submit translations via attachments to BZ
> issues.  They can demonstrate merit and then be voted in as
> committers.
> 
>> So, maybe we let Jürgen deal with Infrastructure about what's possible.
>> 
>> I'll comment on it during the IPMC review on general@i.a.o.
>> 
>> Regards,
>> Dave
>> 
>> 
>>> 
>>> Regards
>>> Ricardo
>> 


Mime
View raw message