New RFR -- Zanata instance

[noriko]

Dimitris Glezos さんは書きました:
> On Tue, Jun 21, 2011 at 11:19 AM, Kévin Raymond
> <shaiton at fedoraproject.org> wrote:
>> On Tue, Jun 21, 2011 at 2:32 AM, Mike McGrath <mmcgrath at redhat.com> wrote:
>>> On Tue, 21 Jun 2011, noriko wrote:
>>>
>>>> hi,
>>>>
>>>> Sorry replying after closing the ticket, I was traveling Japan last two
>>>> weeks and had no internet connection. Now I like you all involved to
>>>> reconsider this RFR.
>>>>
>>>> As Jared pointed, it was clearly agreed to revisit. This offer is great
>>>> opportunity for all translators to test new tool which has potential to
>>>> be hosted at Fedora ground. And as Mike said, the conversation should
>>>> start with translators, so that we will be able to come back. It is best
>>>> to have a test instance within fp.o.
>>>>
>>>> Please notice that this RFR is just asking a test, not a move. As
>>>> someone refer to the need or the cost (people power or $$), without
>>>> testing, translators will not be able to evaluate those whether it is
>>>> worth or not. I like you all to remind that it is not zero cost but it
>>>> does cost running transifex for Fedora at tx.net as well, manpower of
>>>> Infidex.
>>>>
>>>> I have been strong supporter of Transifex, and still am confident that
>>>> Transifex is excellent tool for translators in some projects. However
>>>> there is no reason to deny other tools because it was just
>>>> developed/introduced after Transifex. Most of translators even do not
>>>> notice this discussion. I beg you, please give us a chance and let us
>>>> discuss ourselves before making any decision somewhere we do not know.
>>>> Please do not squash the opportunity for us.
>>>>
>>> They don't notice the discussion because no one has taken it there.  I'm
>>> not sure why we're even still discussing this here since not a single
>>> email has been sent to http://lists.fedoraproject.org/pipermail/trans/ wrt
>>> Zanata.  It's not free for us to just hand over resources that could be
>>> better planned.  So I suggest again, start with the translators.  If
>>> someone on the infrastructure team is willing to sponsor it I'm sure
>>> they'll speak up but it'd be a much better sell if the translation team
>>> actually got together and said "sure, lets take a look"
>> From the translator side, the best would be to introduce the tool
>> (explain workflow/usage differences) before asking them to try it.
> 
> I also think we should have good documentation before requesting time
> to try out a tool -- eg. how does it compare with alternative tools
> such as Transifex, Pootle, Damned Lies, especially in the areas that
> are important to Fedora.

Good point, I believe that the project team should have some 
documentation. But also it can be said that it is hard to discuss 
without seeing. A picture is worth a thousand words.

>> We should not close our mind, but asking the busy infra to work on
>> something that nobody know (from infra as well as translators) is
>> quite heavy.

Yp, thanks for this input! If this is the case I also disagree.
So far, I do not see any manpower requested in this RFR, but I can see 
that there is the project team already formed and it asks for only 
approval from Infra team on consuming certain resources.

>> The best for now should be to take few translators who will install
>> this in their own hardware and try it for few weeks and come back
>> later.

Personally I understand that it would be difficult for many of 
translators to install and test it by themselves. It must be much easier 
and straight forward to provide a sandbox to play around. If the project 
team can not maintain that sandbox just for a test, it would not be 
worth to proceed.

noriko

>> We also need to know the difference in the developer side.
> 
> 
>