You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openoffice.apache.org by Andrea Pescetti <pe...@apache.org> on 2013/03/01 00:51:42 UTC
Re: update service for not released languages [was: Re: Registration]
On 28/02/2013 Oliver-Rainer Wittmann wrote:
> I had a look and an additional text is possible - see attached screenshot.
> The additional text should be translated into the corresponding language.
This could be a good idea. We surely don't want to notify users that an
update is available if it isn't (in their language).
But if we manage to properly communicate in the dialog box that this is
actually a call for volunteers (and write it in the corresponding
language so that it cannot be misinterpreted) this could be a nice way
to involve new people.
Based on the experience we had so far with teams, I would honestly wait
that Pootle is available for all interested languages before we do this.
Or at least we should see if we have a timeframe for a new Pootle setup
that will maximize the likelihood to actually recruit new volunteers.
Ideally, we should also have (unannounced) development builds available
in the target languages to be able to point volunteers to them when they
write.
Regards,
Andrea.
Re: update service for not released languages [was: Re: Registration]
Posted by Rob Weir <ro...@apache.org>.
On Fri, Mar 1, 2013 at 3:00 PM, Andrea Pescetti <pe...@apache.org> wrote:
> Rob Weir wrote:
>>
>> Are we expecting any change to Pootle in the near term? If so, what and
>> when?
>
>
> The changes look trivial but are not so trivial since Pootle accounts are
> now linked to committer accounts. Currently anonymous users can suggest
> translations, and users cannot register (they have to be committers).
>
> We need, for many good reasons (traceability, accountability...) that
> anonymous suggestions are disabled and, at the same time, account creation
> for non-committers is possible.
>
> I was concerned about losing content but apparently Jan reassured here and
> on the l10n list that we are in a position to use/import strings from Pootle
> at this stage. So it is a configuration/policy issue only.
>
>
>> I would not recommend waiting too long. We've shown that offline
>> translation is quite reasonable.. Most of the 3.4.1 languages were
>> done that way.
>
>
> It worked, but the current process is really demanding on the motivation (or
> technical skills) of volunteers. Compare an answer like "Welcome! Next week
> we will send you a link to a .tar.bz2 archive containing 240 PO files that
> you should open individually, translate and send us back" to "Welcome!
> Please register at https://translate.apache.org/ and start translating now".
>
> Then, if you ask me, I would probably prefer the 240 PO files, but the
> majority of new volunteers will be perfectly at home with Pootle. So I
> believe we should fix it before any other mass-recruitment actions.
>
OK. But then maybe let's agree on a date, and if the Pootle service
is not able to handle non-committer users by then, then we go forward
with backup plan for using POEdit and offline translation. It is
better than nothing, and we know it works. But I don't want to see us
in a situation when it is May and we're still waiting for Pootle.
-Rob
> Regards,
> Andrea.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by Rob Weir <ro...@apache.org>.
On Sat, Mar 9, 2013 at 5:28 PM, janI <ja...@apache.org> wrote:
> On 9 March 2013 21:57, Andrea Pescetti <pe...@apache.org> wrote:
>
>> janI wrote:
>>
>>> that is perfectly within the ruleset.
>>>
>>
>> Perfect.
>>
>>
>> I thought the reason for local users was
>>> - get their work named or
>>> - allow them to do review (not possible on suggestions) or
>>> - save committers work
>>> as you suggest I cannot see the difference to using anonymous as we have
>>> it
>>> today ? What did I miss ?
>>>
>>
>> Well, the first item is very important. The step from anonymous
>> contributors to authenticated contributors is a big one: it enables
>> traceability, accountability and accommodates licensing concerns; it allows
>> to evaluate contributions from individual authors. The other improvements
>> are very good, but I see the first one as the key.
>
>
> making support for non-ldap users is no trivial task (just think of spam
>> protection) so clearly this effort should counter weighted of other
>> advantages.
>>
>
> Manual creation/activation of accounts would be fine too: it would still be
>> a significant improvement over what we are doing now.
>
>
> The reference to "accommodates licensing concerns" actually calls for
> another change in the setup, somewhere in the login screen, the user must
> be made aware of or see the ASF license and not standard. I hope we can
> configure it in 2.5.
>
We currently accept patches from non-committers to the mailing list
and to Bugzilla without any special license display. But having the
contributor's contact information helps since if questions arise we
can follow up with them.
Anything short of an ICLA requires review by a committer to make it
into the product. The reviewer should raise any technical concerns,
and licensing ones as well. For example, if we suddenly get a
substantial contribution, code or translation, from someone never
associated with the project before, then the reviewer might clarify
the situation with the contributor. That's why the current setup is
non-optimal -- we don't know who the contributor is.
Regards,
-Rob
> I do not share the opinion that we need to rush on this subject, most
> translations are done offline so if we as an example spent the same time
> improving download/upload it would be much more useful for translators (and
> volunteers who prepare .po files).
>
> I have promised infra to do the upgrade and regular maintenance, which I
> will do (and am doing), but in view of the discussion "support of mwiki
> depending on my person", I do not volunteer to manage/enable account
> creation and thereby create a real dependency (or the feeling of one).
>
> pootle (translate-vm) is a supported infra "product" available to all
> projects, so making a Jira os the correct way to get such changes (with
> manual account creation it is in the installed version via use of the
> backend authentication script). With a Jira we are sure that infra supports
> and maintains the change (including manual creation if needed).
>
> rgds
> jan I.
>
>
>>
>> Regards,
>> Andrea.
>>
>> ------------------------------**------------------------------**---------
>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<de...@openoffice.apache.org>
>> For additional commands, e-mail: dev-help@openoffice.apache.org
>>
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by janI <ja...@apache.org>.
On 9 March 2013 21:57, Andrea Pescetti <pe...@apache.org> wrote:
> janI wrote:
>
>> that is perfectly within the ruleset.
>>
>
> Perfect.
>
>
> I thought the reason for local users was
>> - get their work named or
>> - allow them to do review (not possible on suggestions) or
>> - save committers work
>> as you suggest I cannot see the difference to using anonymous as we have
>> it
>> today ? What did I miss ?
>>
>
> Well, the first item is very important. The step from anonymous
> contributors to authenticated contributors is a big one: it enables
> traceability, accountability and accommodates licensing concerns; it allows
> to evaluate contributions from individual authors. The other improvements
> are very good, but I see the first one as the key.
making support for non-ldap users is no trivial task (just think of spam
> protection) so clearly this effort should counter weighted of other
> advantages.
>
Manual creation/activation of accounts would be fine too: it would still be
> a significant improvement over what we are doing now.
The reference to "accommodates licensing concerns" actually calls for
another change in the setup, somewhere in the login screen, the user must
be made aware of or see the ASF license and not standard. I hope we can
configure it in 2.5.
I do not share the opinion that we need to rush on this subject, most
translations are done offline so if we as an example spent the same time
improving download/upload it would be much more useful for translators (and
volunteers who prepare .po files).
I have promised infra to do the upgrade and regular maintenance, which I
will do (and am doing), but in view of the discussion "support of mwiki
depending on my person", I do not volunteer to manage/enable account
creation and thereby create a real dependency (or the feeling of one).
pootle (translate-vm) is a supported infra "product" available to all
projects, so making a Jira os the correct way to get such changes (with
manual account creation it is in the installed version via use of the
backend authentication script). With a Jira we are sure that infra supports
and maintains the change (including manual creation if needed).
rgds
jan I.
>
> Regards,
> Andrea.
>
> ------------------------------**------------------------------**---------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<de...@openoffice.apache.org>
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
>
Re: update service for not released languages [was: Re: Registration]
Posted by Andrea Pescetti <pe...@apache.org>.
janI wrote:
> that is perfectly within the ruleset.
Perfect.
> I thought the reason for local users was
> - get their work named or
> - allow them to do review (not possible on suggestions) or
> - save committers work
> as you suggest I cannot see the difference to using anonymous as we have it
> today ? What did I miss ?
Well, the first item is very important. The step from anonymous
contributors to authenticated contributors is a big one: it enables
traceability, accountability and accommodates licensing concerns; it
allows to evaluate contributions from individual authors. The other
improvements are very good, but I see the first one as the key.
> making support for non-ldap users is no trivial task (just think of spam
> protection) so clearly this effort should counter weighted of other
> advantages.
Manual creation/activation of accounts would be fine too: it would still
be a significant improvement over what we are doing now.
Regards,
Andrea.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by Dave Fisher <da...@comcast.net>.
On Mar 9, 2013, at 6:57 AM, janI wrote:
> On Mar 9, 2013 3:18 PM, "Andrea Pescetti" <pe...@apache.org> wrote:
>>
>> janI wrote:
>>>
>>> That does not (as I read it) state that we can bypass RTC for
>>> non-committers. Allowing non-committers access is one thing, but allowing
>>> them to change the source (in this case text) directly is quite another.
>>
>>
>> Sure. The setting for new volunteers would be:
>>
>> 1) No paperwork, no ICLA, just create an account on Pootle
>>
>> 2) Translate through suggestions (equivalent to contributing patches)
>>
>> 3) RTC in place (i.e., strings are committed after a review by a
> committer, exactly as it happens now; what a "review" is in this context
> will vary, as it is now, depending on availability and skills of
> volunteers).
>
> that is perfectly within the ruleset.
>
> I thought the reason for local users was
> - get their work named or
> - allow them to do review (not possible on suggestions) or
> - save committers work
>
> as you suggest I cannot see the difference to using anonymous as we have it
> today ? What did I miss ?
We should know who is making the suggestion. If someone makes a large number of high quality suggestions then they are a good candidate for project committer.
If suggestions are all anonymous then we haven't a clue.
By naming the work then items 2 and 3 are future enhancements as more translators become committers.
>
> making support for non-ldap users is no trivial task (just think of spam
> protection) so clearly this effort should counter weighted of other
> advantages.
True, we should not be obfuscating work by moving it elsewhere.
Regards,
Dave
>
> rgds
> jan i
>>
>> Regards,
>> Andrea.
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
>> For additional commands, e-mail: dev-help@openoffice.apache.org
>>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by janI <ja...@apache.org>.
On Mar 9, 2013 3:18 PM, "Andrea Pescetti" <pe...@apache.org> wrote:
>
> janI wrote:
>>
>> That does not (as I read it) state that we can bypass RTC for
>> non-committers. Allowing non-committers access is one thing, but allowing
>> them to change the source (in this case text) directly is quite another.
>
>
> Sure. The setting for new volunteers would be:
>
> 1) No paperwork, no ICLA, just create an account on Pootle
>
> 2) Translate through suggestions (equivalent to contributing patches)
>
> 3) RTC in place (i.e., strings are committed after a review by a
committer, exactly as it happens now; what a "review" is in this context
will vary, as it is now, depending on availability and skills of
volunteers).
that is perfectly within the ruleset.
I thought the reason for local users was
- get their work named or
- allow them to do review (not possible on suggestions) or
- save committers work
as you suggest I cannot see the difference to using anonymous as we have it
today ? What did I miss ?
making support for non-ldap users is no trivial task (just think of spam
protection) so clearly this effort should counter weighted of other
advantages.
rgds
jan i
>
> Regards,
> Andrea.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
Re: update service for not released languages [was: Re: Registration]
Posted by Andrea Pescetti <pe...@apache.org>.
janI wrote:
> That does not (as I read it) state that we can bypass RTC for
> non-committers. Allowing non-committers access is one thing, but allowing
> them to change the source (in this case text) directly is quite another.
Sure. The setting for new volunteers would be:
1) No paperwork, no ICLA, just create an account on Pootle
2) Translate through suggestions (equivalent to contributing patches)
3) RTC in place (i.e., strings are committed after a review by a
committer, exactly as it happens now; what a "review" is in this context
will vary, as it is now, depending on availability and skills of
volunteers).
Regards,
Andrea.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by janI <ja...@apache.org>.
On 9 March 2013 12:52, Andrea Pescetti <pe...@apache.org> wrote:
> On 03/03/2013 janI wrote:
>
>> On 3 March 2013 17:47, Andrea Pescetti wrote:
>>
>>> 1) Check on the Pootle 2.5 release date and features.
>>>
>> I would like to see it running on other sites the translate itself, but I
>> am just a negative (have been too long in support). My rule of thumb is
>> release date + 1 month, in order not to fight fight with start problems.
>>
>
> Here I agree with Rob that we need to set a deadline. A natural one is the
> translation period set at https://cwiki.apache.org/**
> confluence/display/OOOUSERS/**AOO+4.0+Release+Planning<https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Release+Planning>(beginning in one month). So, considering installation and testing, we
> would need Pootle 2.5 to be available soon. I've asked the developers if
> they have a timeline, just to get an idea.
>
> 2) Check that policy-wise it's fine to authenticate committers on LDAP and
>>> all other volunteers on a local backend. ...
>>>
>> infra (gmcdonald) was not positive, but I still think we
>> have a case and should go for it...I do however think a compromise could
>> be
>> a signed ICLA.
>>
>
> This has been clarified in the meantime on the Incubator lists (in a
> discussion otherwise unrelated to OpenOffice). No ICLA needed.
> http://mail-archives.apache.**org/mod_mbox/incubator-**
> general/201303.mbox/%3CCAAS6%**3D7hybut%**3DLGZQRkuuJPXKK4KPS6CiXDYE5-**
> QTmvguYHOVFA%40mail.gmail.com%**3E<http://mail-archives.apache.org/mod_mbox/incubator-general/201303.mbox/%3CCAAS6%3D7hybut%3DLGZQRkuuJPXKK4KPS6CiXDYE5-QTmvguYHOVFA%40mail.gmail.com%3E>
>
That does not (as I read it) state that we can bypass RTC for
non-committers. Allowing non-committers access is one thing, but allowing
them to change the source (in this case text) directly is quite another.
Translated text is being compiled into our binaries, so there are no
difference for source code and and translations.
I have no problem (infra might see it differently) if non-committers do a
login and provide suggestions (as today), that way we keep RTC.
rgds
Jan I.
>
> 3) Optimize performance so that Pootle is actually usable by several
>>> users.
>>>
>> That is something on my list of todos, and infra ask me regulary when I do
>> it. ...a bottle of good italian wine when if finally works,
>> together with genLang.
>>
>
> OK. And OK for the bottle too!
>
> Regards,
> Andrea.
>
> ------------------------------**------------------------------**---------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<de...@openoffice.apache.org>
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
>
Re: update service for not released languages [was: Re: Registration]
Posted by Jürgen Schmidt <jo...@gmail.com>.
On 4/4/13 8:26 AM, janI wrote:
> On 4 April 2013 06:19, Juergen Schmidt <jo...@gmail.com> wrote:
>
>> Am Mittwoch, 3. April 2013 um 22:46 schrieb Rob Weir:
>>> On Sat, Mar 9, 2013 at 6:52 AM, Andrea Pescetti <pe...@apache.org>
>> wrote:
>>>
>>>> On 03/03/2013 janI wrote:
>>>>
>>>>> On 3 March 2013 17:47, Andrea Pescetti wrote:
>>>>>
>>>>>> 1) Check on the Pootle 2.5 release date and features.
>>>>> I would like to see it running on other sites the translate itself,
>> but I
>>>>>
>>>>> am just a negative (have been too long in support). My rule of thumb
>> is
>>>>> release date + 1 month, in order not to fight fight with start
>> problems.
>>>>>
>>>>
>>>>
>>>> Here I agree with Rob that we need to set a deadline. A natural one is
>> the
>>>> translation period set at https://cwiki.apache.org/**
>>>> confluence/display/OOOUSERS/**AOO+4.0+Release+Planning<
>> https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Release+Planning>(beginning
>> in one month). So, considering installation and testing, we
>>>> would need Pootle 2.5 to be available soon. I've asked the developers
>> if
>>>> they have a timeline, just to get an idea.
>>>>
>>>
>>> Almost a month has passed since I made this proposal. Do we have a sense
>>> now whether the Pootle upgrade will occur in time for AOO 4.0
>>> translations? Or I should I go ahead with the a call for translations for
>>> AOO 4.0 using POEdit and similar off-line tools? This would include
>>> reaching out to users via the update notification mechanism.
>>>
>>>
>>
>> I will try to figure it out today...
>> But it is to early and we don't have the new translation files in place
>> yet. We have to merge the sidebar branch first and that is already in
>> preparation.
>>
>> 2.5 is not released and will most problaly not make it before we need it.
> BUT the translation interface have not changed, it is inner working and
> download possibilities.
>
> I think we need offline tools, I would not to mass translation on the
> server.
>
> To me the bigger question is, do we start translation with our current file
> structure, or will genLang be in place (heavely reduced file count).
> Integration into build system has been a lot more difficult than i
> expected, especially for helpcontent2. I have concentrated on helpcontent2,
> and are very close with that, so I expect us to take a discussion/vote next
> week, whether or not to integrate it in trunk.
Hi Jan,
having the new tooling in place would be great but we don't have
pressure. We should take all the time that we need to make it final and
complete and well tested. Before we integrate it we should test your
branch with some languages in detail on at least 3 platforms (Linux,
windows, MacOS).
You did a fantastic job and it will be a huge step forward. We can
integrate it probably at any time when it is ready.
On the other hand it is a very important part of our build system and my
experience told me that we should test it deeply. Don't get me wrong I
am sure you made a good job here, it's simply that I have seen too often
that we got many problems with less.
Don't put yourself under pressure.
I can build your branch on MacOS. What do I have to do exactly to
trigger your new tooling? I will ping you on IRC ...
Juergen
>
> rgds
> Jan I
>
>> Juergen
>>>
>>> -Rob
>>>
>>>
>>>
>>>
>>>> 2) Check that policy-wise it's fine to authenticate committers on LDAP
>> and
>>>>>> all other volunteers on a local backend. ...
>>>>>
>>>>> infra (gmcdonald) was not positive, but I still think we
>>>>> have a case and should go for it...I do however think a compromise
>> could
>>>>> be
>>>>> a signed ICLA.
>>>>>
>>>>
>>>>
>>>> This has been clarified in the meantime on the Incubator lists (in a
>>>> discussion otherwise unrelated to OpenOffice). No ICLA needed.
>>>> http://mail-archives.apache.**org/mod_mbox/incubator-**
>>>>
>> general/201303.mbox/%3CCAAS6%**3D7hybut%**3DLGZQRkuuJPXKK4KPS6CiXDYE5-**
>>>> QTmvguYHOVFA%40mail.gmail.com%**3E<
>> http://mail-archives.apache.org/mod_mbox/incubator-general/201303.mbox/%3CCAAS6%3D7hybut%3DLGZQRkuuJPXKK4KPS6CiXDYE5-QTmvguYHOVFA%40mail.gmail.com%3E
>>>
>>>>
>>>> 3) Optimize performance so that Pootle is actually usable by several
>>>>>> users.
>>>>>
>>>>> That is something on my list of todos, and infra ask me regulary
>> when I do
>>>>> it. ...a bottle of good italian wine when if finally works,
>>>>> together with genLang.
>>>>>
>>>>
>>>>
>>>> OK. And OK for the bottle too!
>>>>
>>>> Regards,
>>>> Andrea.
>>>>
>>>>
>> ------------------------------**------------------------------**---------
>>>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<
>> dev-unsubscribe@openoffice.apache.org>
>>>> For additional commands, e-mail: dev-help@openoffice.apache.org
>>>>
>>>
>>>
>>>
>>
>>
>>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by janI <ja...@apache.org>.
On 4 April 2013 06:19, Juergen Schmidt <jo...@gmail.com> wrote:
> Am Mittwoch, 3. April 2013 um 22:46 schrieb Rob Weir:
> > On Sat, Mar 9, 2013 at 6:52 AM, Andrea Pescetti <pe...@apache.org>
> wrote:
> >
> > > On 03/03/2013 janI wrote:
> > >
> > > > On 3 March 2013 17:47, Andrea Pescetti wrote:
> > > >
> > > > > 1) Check on the Pootle 2.5 release date and features.
> > > > I would like to see it running on other sites the translate itself,
> but I
> > > >
> > > > am just a negative (have been too long in support). My rule of thumb
> is
> > > > release date + 1 month, in order not to fight fight with start
> problems.
> > > >
> > >
> > >
> > > Here I agree with Rob that we need to set a deadline. A natural one is
> the
> > > translation period set at https://cwiki.apache.org/**
> > > confluence/display/OOOUSERS/**AOO+4.0+Release+Planning<
> https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Release+Planning>(beginning
> in one month). So, considering installation and testing, we
> > > would need Pootle 2.5 to be available soon. I've asked the developers
> if
> > > they have a timeline, just to get an idea.
> > >
> >
> > Almost a month has passed since I made this proposal. Do we have a sense
> > now whether the Pootle upgrade will occur in time for AOO 4.0
> > translations? Or I should I go ahead with the a call for translations for
> > AOO 4.0 using POEdit and similar off-line tools? This would include
> > reaching out to users via the update notification mechanism.
> >
> >
>
> I will try to figure it out today...
> But it is to early and we don't have the new translation files in place
> yet. We have to merge the sidebar branch first and that is already in
> preparation.
>
> 2.5 is not released and will most problaly not make it before we need it.
BUT the translation interface have not changed, it is inner working and
download possibilities.
I think we need offline tools, I would not to mass translation on the
server.
To me the bigger question is, do we start translation with our current file
structure, or will genLang be in place (heavely reduced file count).
Integration into build system has been a lot more difficult than i
expected, especially for helpcontent2. I have concentrated on helpcontent2,
and are very close with that, so I expect us to take a discussion/vote next
week, whether or not to integrate it in trunk.
rgds
Jan I
> Juergen
> >
> > -Rob
> >
> >
> >
> >
> > > 2) Check that policy-wise it's fine to authenticate committers on LDAP
> and
> > > > > all other volunteers on a local backend. ...
> > > >
> > > > infra (gmcdonald) was not positive, but I still think we
> > > > have a case and should go for it...I do however think a compromise
> could
> > > > be
> > > > a signed ICLA.
> > > >
> > >
> > >
> > > This has been clarified in the meantime on the Incubator lists (in a
> > > discussion otherwise unrelated to OpenOffice). No ICLA needed.
> > > http://mail-archives.apache.**org/mod_mbox/incubator-**
> > >
> general/201303.mbox/%3CCAAS6%**3D7hybut%**3DLGZQRkuuJPXKK4KPS6CiXDYE5-**
> > > QTmvguYHOVFA%40mail.gmail.com%**3E<
> http://mail-archives.apache.org/mod_mbox/incubator-general/201303.mbox/%3CCAAS6%3D7hybut%3DLGZQRkuuJPXKK4KPS6CiXDYE5-QTmvguYHOVFA%40mail.gmail.com%3E
> >
> > >
> > > 3) Optimize performance so that Pootle is actually usable by several
> > > > > users.
> > > >
> > > > That is something on my list of todos, and infra ask me regulary
> when I do
> > > > it. ...a bottle of good italian wine when if finally works,
> > > > together with genLang.
> > > >
> > >
> > >
> > > OK. And OK for the bottle too!
> > >
> > > Regards,
> > > Andrea.
> > >
> > >
> ------------------------------**------------------------------**---------
> > > To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<
> dev-unsubscribe@openoffice.apache.org>
> > > For additional commands, e-mail: dev-help@openoffice.apache.org
> > >
> >
> >
> >
>
>
>
Re: update service for not released languages [was: Re:
Registration]
Posted by Juergen Schmidt <jo...@gmail.com>.
Am Mittwoch, 3. April 2013 um 22:46 schrieb Rob Weir:
> On Sat, Mar 9, 2013 at 6:52 AM, Andrea Pescetti <pe...@apache.org> wrote:
>
> > On 03/03/2013 janI wrote:
> >
> > > On 3 March 2013 17:47, Andrea Pescetti wrote:
> > >
> > > > 1) Check on the Pootle 2.5 release date and features.
> > > I would like to see it running on other sites the translate itself, but I
> > >
> > > am just a negative (have been too long in support). My rule of thumb is
> > > release date + 1 month, in order not to fight fight with start problems.
> > >
> >
> >
> > Here I agree with Rob that we need to set a deadline. A natural one is the
> > translation period set at https://cwiki.apache.org/**
> > confluence/display/OOOUSERS/**AOO+4.0+Release+Planning<https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Release+Planning>(beginning in one month). So, considering installation and testing, we
> > would need Pootle 2.5 to be available soon. I've asked the developers if
> > they have a timeline, just to get an idea.
> >
>
> Almost a month has passed since I made this proposal. Do we have a sense
> now whether the Pootle upgrade will occur in time for AOO 4.0
> translations? Or I should I go ahead with the a call for translations for
> AOO 4.0 using POEdit and similar off-line tools? This would include
> reaching out to users via the update notification mechanism.
>
>
I will try to figure it out today...
But it is to early and we don't have the new translation files in place yet. We have to merge the sidebar branch first and that is already in preparation.
Juergen
>
> -Rob
>
>
>
>
> > 2) Check that policy-wise it's fine to authenticate committers on LDAP and
> > > > all other volunteers on a local backend. ...
> > >
> > > infra (gmcdonald) was not positive, but I still think we
> > > have a case and should go for it...I do however think a compromise could
> > > be
> > > a signed ICLA.
> > >
> >
> >
> > This has been clarified in the meantime on the Incubator lists (in a
> > discussion otherwise unrelated to OpenOffice). No ICLA needed.
> > http://mail-archives.apache.**org/mod_mbox/incubator-**
> > general/201303.mbox/%3CCAAS6%**3D7hybut%**3DLGZQRkuuJPXKK4KPS6CiXDYE5-**
> > QTmvguYHOVFA%40mail.gmail.com%**3E<http://mail-archives.apache.org/mod_mbox/incubator-general/201303.mbox/%3CCAAS6%3D7hybut%3DLGZQRkuuJPXKK4KPS6CiXDYE5-QTmvguYHOVFA%40mail.gmail.com%3E>
> >
> > 3) Optimize performance so that Pootle is actually usable by several
> > > > users.
> > >
> > > That is something on my list of todos, and infra ask me regulary when I do
> > > it. ...a bottle of good italian wine when if finally works,
> > > together with genLang.
> > >
> >
> >
> > OK. And OK for the bottle too!
> >
> > Regards,
> > Andrea.
> >
> > ------------------------------**------------------------------**---------
> > To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<de...@openoffice.apache.org>
> > For additional commands, e-mail: dev-help@openoffice.apache.org
> >
>
>
>
Re: update service for not released languages [was: Re: Registration]
Posted by Rob Weir <ro...@apache.org>.
On Sat, Mar 9, 2013 at 6:52 AM, Andrea Pescetti <pe...@apache.org> wrote:
> On 03/03/2013 janI wrote:
>
>> On 3 March 2013 17:47, Andrea Pescetti wrote:
>>
>>> 1) Check on the Pootle 2.5 release date and features.
>>>
>> I would like to see it running on other sites the translate itself, but I
>>
>> am just a negative (have been too long in support). My rule of thumb is
>> release date + 1 month, in order not to fight fight with start problems.
>>
>
> Here I agree with Rob that we need to set a deadline. A natural one is the
> translation period set at https://cwiki.apache.org/**
> confluence/display/OOOUSERS/**AOO+4.0+Release+Planning<https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Release+Planning>(beginning in one month). So, considering installation and testing, we
> would need Pootle 2.5 to be available soon. I've asked the developers if
> they have a timeline, just to get an idea.
>
>
Almost a month has passed since I made this proposal. Do we have a sense
now whether the Pootle upgrade will occur in time for AOO 4.0
translations? Or I should I go ahead with the a call for translations for
AOO 4.0 using POEdit and similar off-line tools? This would include
reaching out to users via the update notification mechanism.
-Rob
> 2) Check that policy-wise it's fine to authenticate committers on LDAP and
>>> all other volunteers on a local backend. ...
>>>
>> infra (gmcdonald) was not positive, but I still think we
>> have a case and should go for it...I do however think a compromise could
>> be
>> a signed ICLA.
>>
>
> This has been clarified in the meantime on the Incubator lists (in a
> discussion otherwise unrelated to OpenOffice). No ICLA needed.
> http://mail-archives.apache.**org/mod_mbox/incubator-**
> general/201303.mbox/%3CCAAS6%**3D7hybut%**3DLGZQRkuuJPXKK4KPS6CiXDYE5-**
> QTmvguYHOVFA%40mail.gmail.com%**3E<http://mail-archives.apache.org/mod_mbox/incubator-general/201303.mbox/%3CCAAS6%3D7hybut%3DLGZQRkuuJPXKK4KPS6CiXDYE5-QTmvguYHOVFA%40mail.gmail.com%3E>
>
> 3) Optimize performance so that Pootle is actually usable by several
>>> users.
>>>
>> That is something on my list of todos, and infra ask me regulary when I do
>> it. ...a bottle of good italian wine when if finally works,
>> together with genLang.
>>
>
> OK. And OK for the bottle too!
>
> Regards,
> Andrea.
>
> ------------------------------**------------------------------**---------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.**apache.org<de...@openoffice.apache.org>
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
>
Re: update service for not released languages [was: Re: Registration]
Posted by Andrea Pescetti <pe...@apache.org>.
On 03/03/2013 janI wrote:
> On 3 March 2013 17:47, Andrea Pescetti wrote:
>> 1) Check on the Pootle 2.5 release date and features.
> I would like to see it running on other sites the translate itself, but I
> am just a negative (have been too long in support). My rule of thumb is
> release date + 1 month, in order not to fight fight with start problems.
Here I agree with Rob that we need to set a deadline. A natural one is
the translation period set at
https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Release+Planning
(beginning in one month). So, considering installation and testing, we
would need Pootle 2.5 to be available soon. I've asked the developers if
they have a timeline, just to get an idea.
>> 2) Check that policy-wise it's fine to authenticate committers on LDAP and
>> all other volunteers on a local backend. ...
> infra (gmcdonald) was not positive, but I still think we
> have a case and should go for it...I do however think a compromise could be
> a signed ICLA.
This has been clarified in the meantime on the Incubator lists (in a
discussion otherwise unrelated to OpenOffice). No ICLA needed.
http://mail-archives.apache.org/mod_mbox/incubator-general/201303.mbox/%3CCAAS6%3D7hybut%3DLGZQRkuuJPXKK4KPS6CiXDYE5-QTmvguYHOVFA%40mail.gmail.com%3E
>> 3) Optimize performance so that Pootle is actually usable by several users.
> That is something on my list of todos, and infra ask me regulary when I do
> it. ...a bottle of good italian wine when if finally works,
> together with genLang.
OK. And OK for the bottle too!
Regards,
Andrea.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: update service for not released languages [was: Re: Registration]
Posted by janI <ja...@apache.org>.
On 3 March 2013 17:47, Andrea Pescetti <pe...@apache.org> wrote:
> On 01/03/2013 janI wrote:
>
>> We need to wait for the release of pootle, last I checked it was still not
>> official. Once 2.5 is released I will update translate-vm.
>> Please consider the translate-vm, is currently NOT configured for a higher
>> online volume (about 3 users and mysql is strugling). Infra has agreed
>> that
>> I do performance tuning after installing the new release.
>> It should be possible to configure 2.5 to use both ldap and local db.
>>
>
> OK, thanks for the detailed explanations. So the steps would be:
>
> 1) Check on the Pootle 2.5 release date and features. Indeed I can't find
> information easily at http://translate.sourceforge.**net/<http://translate.sourceforge.net/>so we will have to bother the developers. If you need any help in
> contacting them, just ask.
>
I would like to see it running on other sites the translate itself, but I
am just a negative (have been too long in support). My rule of thumb is
release date + 1 month, in order not to fight fight with start problems.
> 2) Check that policy-wise it's fine to authenticate committers on LDAP and
> all other volunteers on a local backend. I believe it's fine (and actually
> very helpful for our use case), but I'm prepared to raise the issue
> properly at due time.
>
You might need to, infra (gmcdonald) was not positive, but I still think we
have a case and should go for it...I do however think a compromise could be
a signed ICLA.
>
> 3) Optimize performance so that Pootle is actually usable by several users.
>
That is something on my list of todos, and infra ask me regulary when I do
it.
>
> Am I missing any other steps?
>
Not really, apart from a bottle of good italian wine when if finally works,
together with genLang.
>
> Regards,
> Andrea.
>
Re: update service for not released languages [was: Re: Registration]
Posted by Andrea Pescetti <pe...@apache.org>.
On 01/03/2013 janI wrote:
> We need to wait for the release of pootle, last I checked it was still not
> official. Once 2.5 is released I will update translate-vm.
> Please consider the translate-vm, is currently NOT configured for a higher
> online volume (about 3 users and mysql is strugling). Infra has agreed that
> I do performance tuning after installing the new release.
> It should be possible to configure 2.5 to use both ldap and local db.
OK, thanks for the detailed explanations. So the steps would be:
1) Check on the Pootle 2.5 release date and features. Indeed I can't
find information easily at http://translate.sourceforge.net/ so we will
have to bother the developers. If you need any help in contacting them,
just ask.
2) Check that policy-wise it's fine to authenticate committers on LDAP
and all other volunteers on a local backend. I believe it's fine (and
actually very helpful for our use case), but I'm prepared to raise the
issue properly at due time.
3) Optimize performance so that Pootle is actually usable by several users.
Am I missing any other steps?
Regards,
Andrea.
Re: update service for not released languages [was: Re: Registration]
Posted by Rob Weir <ro...@apache.org>.
On Fri, Mar 1, 2013 at 3:26 PM, janI <ja...@apache.org> wrote:
> On 1 March 2013 21:00, Andrea Pescetti <pe...@apache.org> wrote:
>
>> Rob Weir wrote:
>>
>>> Are we expecting any change to Pootle in the near term? If so, what and
>>> when?
>>>
>>
>> The changes look trivial but are not so trivial since Pootle accounts are
>> now linked to committer accounts. Currently anonymous users can suggest
>> translations, and users cannot register (they have to be committers).
>>
>> We need, for many good reasons (traceability, accountability...) that
>> anonymous suggestions are disabled and, at the same time, account creation
>> for non-committers is possible.
>>
>> I was concerned about losing content but apparently Jan reassured here and
>> on the l10n list that we are in a position to use/import strings from
>> Pootle at this stage. So it is a configuration/policy issue only.
>
>
> The import might be partly manual or a "perl/python" volunteer writes a
> small script. It is quite simple, match source file, english text in the
> translated po file, with the new po file, and update (I can write the req.
> but I am not fluent in perl/python).
>
> Offering an import possibility was a demand, since I cannot (and will not)
> request a full stop on translation.
>
>
>>
>> I would not recommend waiting too long. We've shown that offline
>>> translation is quite reasonable.. Most of the 3.4.1 languages were
>>> done that way.
>>>
>>
>> It worked, but the current process is really demanding on the motivation
>> (or technical skills) of volunteers. Compare an answer like "Welcome! Next
>> week we will send you a link to a .tar.bz2 archive containing 240 PO files
>> that you should open individually, translate and send us back" to "Welcome!
>> Please register at https://translate.apache.org/ and start translating
>> now".
>>
>> Then, if you ask me, I would probably prefer the 240 PO files, but the
>> majority of new volunteers will be perfectly at home with Pootle. So I
>> believe we should fix it before any other mass-recruitment actions.
>>
>
> We need to wait for the release of pootle, last I checked it was still not
> official. Once 2.5 is released I will update translate-vm.
>
> Please consider the translate-vm, is currently NOT configured for a higher
> online volume (about 3 users and mysql is strugling). Infra has agreed that
> I do performance tuning after installing the new release.
>
> It should be possible to configure 2.5 to use both ldap and local db. The
> current version is either/or making it technically impossible to allow
> non-commiter login. Since pootle is seen as a asf-wide service, we need to
> get the acceptance from infra.
>
Ah. OK. That is the part I did not know before. This is not
entirely an Infra policy thing, but we're waiting for some new
technical capabilities in Pootle 2.5.
-Rob
> Please remember genLang will reduce the number of files to 54 files (1 pr
> module, and for helpcontent2 1 pr sub directory). The extraction part of
> genLang is nearly ready for production (as you might have seen in the
> commits).
>
> rgds
> jan I.
>
> Regards,
>> Andrea.
>>
Re: update service for not released languages [was: Re: Registration]
Posted by janI <ja...@apache.org>.
On 1 March 2013 21:00, Andrea Pescetti <pe...@apache.org> wrote:
> Rob Weir wrote:
>
>> Are we expecting any change to Pootle in the near term? If so, what and
>> when?
>>
>
> The changes look trivial but are not so trivial since Pootle accounts are
> now linked to committer accounts. Currently anonymous users can suggest
> translations, and users cannot register (they have to be committers).
>
> We need, for many good reasons (traceability, accountability...) that
> anonymous suggestions are disabled and, at the same time, account creation
> for non-committers is possible.
>
> I was concerned about losing content but apparently Jan reassured here and
> on the l10n list that we are in a position to use/import strings from
> Pootle at this stage. So it is a configuration/policy issue only.
The import might be partly manual or a "perl/python" volunteer writes a
small script. It is quite simple, match source file, english text in the
translated po file, with the new po file, and update (I can write the req.
but I am not fluent in perl/python).
Offering an import possibility was a demand, since I cannot (and will not)
request a full stop on translation.
>
> I would not recommend waiting too long. We've shown that offline
>> translation is quite reasonable.. Most of the 3.4.1 languages were
>> done that way.
>>
>
> It worked, but the current process is really demanding on the motivation
> (or technical skills) of volunteers. Compare an answer like "Welcome! Next
> week we will send you a link to a .tar.bz2 archive containing 240 PO files
> that you should open individually, translate and send us back" to "Welcome!
> Please register at https://translate.apache.org/ and start translating
> now".
>
> Then, if you ask me, I would probably prefer the 240 PO files, but the
> majority of new volunteers will be perfectly at home with Pootle. So I
> believe we should fix it before any other mass-recruitment actions.
>
We need to wait for the release of pootle, last I checked it was still not
official. Once 2.5 is released I will update translate-vm.
Please consider the translate-vm, is currently NOT configured for a higher
online volume (about 3 users and mysql is strugling). Infra has agreed that
I do performance tuning after installing the new release.
It should be possible to configure 2.5 to use both ldap and local db. The
current version is either/or making it technically impossible to allow
non-commiter login. Since pootle is seen as a asf-wide service, we need to
get the acceptance from infra.
Please remember genLang will reduce the number of files to 54 files (1 pr
module, and for helpcontent2 1 pr sub directory). The extraction part of
genLang is nearly ready for production (as you might have seen in the
commits).
rgds
jan I.
Regards,
> Andrea.
>
Re: update service for not released languages [was: Re: Registration]
Posted by Andrea Pescetti <pe...@apache.org>.
Rob Weir wrote:
> Are we expecting any change to Pootle in the near term? If so, what and when?
The changes look trivial but are not so trivial since Pootle accounts
are now linked to committer accounts. Currently anonymous users can
suggest translations, and users cannot register (they have to be
committers).
We need, for many good reasons (traceability, accountability...) that
anonymous suggestions are disabled and, at the same time, account
creation for non-committers is possible.
I was concerned about losing content but apparently Jan reassured here
and on the l10n list that we are in a position to use/import strings
from Pootle at this stage. So it is a configuration/policy issue only.
> I would not recommend waiting too long. We've shown that offline
> translation is quite reasonable.. Most of the 3.4.1 languages were
> done that way.
It worked, but the current process is really demanding on the motivation
(or technical skills) of volunteers. Compare an answer like "Welcome!
Next week we will send you a link to a .tar.bz2 archive containing 240
PO files that you should open individually, translate and send us back"
to "Welcome! Please register at https://translate.apache.org/ and start
translating now".
Then, if you ask me, I would probably prefer the 240 PO files, but the
majority of new volunteers will be perfectly at home with Pootle. So I
believe we should fix it before any other mass-recruitment actions.
Regards,
Andrea.
Re: update service for not released languages [was: Re: Registration]
Posted by Rob Weir <ra...@gmail.com>.
On Feb 28, 2013, at 6:52 PM, Andrea Pescetti <pe...@apache.org> wrote:
> On 28/02/2013 Oliver-Rainer Wittmann wrote:
>> I had a look and an additional text is possible - see attached screenshot.
>> The additional text should be translated into the corresponding language.
>
> This could be a good idea. We surely don't want to notify users that an update is available if it isn't (in their language).
>
I think it is still worth a notification, especially in bilingual
areas. A user might prefer to have a more recent release in a 2nd
language than an old release in a 1st language. It is a reasonable
choice either way.
> But if we manage to properly communicate in the dialog box that this is actually a call for volunteers (and write it in the corresponding language so that it cannot be misinterpreted) this could be a nice way to involve new people.
>
> Based on the experience we had so far with teams, I would honestly wait that Pootle is available for all interested languages before we do this. Or at least we should see if we have a timeframe for a new Pootle setup that will maximize the likelihood to actually recruit new volunteers. Ideally, we should also have (unannounced) development builds available in the target languages to be able to point volunteers to them when they write.
>
Are we expecting any change to Pootle in the near term? If so, what and when?
I would not recommend waiting too long. We've shown that offline
translation is quite reasonable.. Most of the 3.4.1 languages were
done that way.
-Rob
> Regards,
> Andrea.
Re: update service for not released languages [was: Re: Registration]
Posted by Oliver-Rainer Wittmann <or...@googlemail.com>.
Hi,
On 01.03.2013 00:51, Andrea Pescetti wrote:
> On 28/02/2013 Oliver-Rainer Wittmann wrote:
>> I had a look and an additional text is possible - see attached
>> screenshot.
>> The additional text should be translated into the corresponding language.
>
> This could be a good idea. We surely don't want to notify users that an
> update is available if it isn't (in their language).
>
> But if we manage to properly communicate in the dialog box that this is
> actually a call for volunteers (and write it in the corresponding
> language so that it cannot be misinterpreted) this could be a nice way
> to involve new people.
>
Yes, that was exactly the purpose why I had a look.
+1 from my side to include such an explaining text in the update
notification dialog.
> Based on the experience we had so far with teams, I would honestly wait
> that Pootle is available for all interested languages before we do this.
> Or at least we should see if we have a timeframe for a new Pootle setup
> that will maximize the likelihood to actually recruit new volunteers.
> Ideally, we should also have (unannounced) development builds available
> in the target languages to be able to point volunteers to them when they
> write.
>
Again +1
Our tools should be ready before we start this promotion.
Best regards, Oliver.