Re: JOSM Translations

2017-02-19 Thread Dirk Stöcker

On Sun, 19 Feb 2017, Martin Koppenhoefer wrote:


I admit I have copied the initial message from Dirk to talk, thinking that
indeed there might be people around of those languages with missing
translations, not knowing that help could be needed but willing to give a
hand:
https://lists.openstreetmap.org/pipermail/talk/2017-February/077579.html

I excuse if this was premature...


I'm not subscribed to any other list anymore, so feel free to spread the 
word. Everthing which helps to improve the situation is welcome.


Sanitizing my language I have seen :-)

Ciao
--
http://www.dstoecker.eu/ (PGP key available)



Re: JOSM Translations

2017-02-19 Thread Dirk Stöcker

On Sun, 19 Feb 2017, Holger Mappt wrote:

It would be interesting to know how important translations are. Maybe it is 
okay for most of the users to use the English version.


The last days:

8079 (31.7%) en
5489 (21.5%) de
2615 (10.3%) ru
2157 ( 8.5%) fr
1541 ( 6.0%) en_GB
1018 ( 4.0%) es
 849 ( 3.3%) it
 709 ( 2.8%) pl
 468 ( 1.8%) ja
 426 ( 1.7%) nl
 311 ( 1.2%) zh_TW
 309 ( 1.2%) pt_BR
 239 ( 0.9%) zh_CN
 223 ( 0.9%) cs
 186 ( 0.7%) en_AU
 175 ( 0.7%) hu
 126 ( 0.5%) sv
 101 ( 0.4%) id
  92 ( 0.4%) uk
  81 ( 0.3%) sk
  57 ( 0.2%) da
  49 ( 0.2%) fi
  48 ( 0.2%) pt
  45 ( 0.2%) ca
  39 ( 0.2%) lt
  31 ( 0.1%) el
  17 ( 0.1%) nb
   5 ( 0.0%) be
   3 ( 0.0%) tr
   3 ( 0.0%) bg
   2 ( 0.0%) et
   2 ( 0.0%) gl
   1 ( 0.0%) vi
   1 ( 0.0%) ast

62% use a non-english version. That number would probably be higher if the 
translations in some languages would be better.


I think one of the issues is the enormous number of strings. We reached 1 
a few weeks ago. And there is a pretty high update frequency. You need days 
to catch up if no translation was done for a few months.


Yes. That's obvious. That burden can only be reduced when distributed to 
more shoulders (althought single people did translate thousands of texts).


Here everybody with more than 3000 translations:
14921 - pitort - ca   = 7959 - ca@valencia= 6962
10844 - andygol- ru   =  290 - uk   =10554
10486 - Viktar Palsciuk- be   =10486
10451 - mmyfl  - zh_CN=10451
10195 - Dirk Stöcker   - de   = 5146 (others: many languages as fixes)
10080 - aceman444  - cs   = 5220 - sk   = 4860
 8594 - Minh Nguyễn- vi   = 8594
 8132 - DiGro  - nl   = 8132
 7543 - edrux  - ast  = 7524 - es   =   19
 7512 - Davide Prade   - it   = 7512
 7441 - RPome  - pt   = 7441
 6726 - fujimoto   - ja   = 6726
 6186 - Casiope
 5972 - Chao-Hsiung Liao   - zh_TW= 5972
 5698 - Jørn   - da   = 5698
 5694 - Dalibor Jelínek- cs   = 5693 - sk   =1
 5656 - Báthory Péter  - hu   = 5656
 5633 - Casiope- fr   = 5633
 5335 - Felipe L.  - pt_BR= 5335
 4710 - Vlado  - sk   = 4710
 4562 - Dee-earlsoft   - en_GB= 4562
 3873 - Emilio Gomez Fernandez - es   = 3872 - fr   =1
 3858 - Sophea Sok - km   = 3858
 3720 - Aleksey Kabanov- ru   = 3720
 3584 - Robert Readman - en_AU= 3266 - en_GB=  318
 3309 - Ettore Atalan  - de   = 3309
 3110 - Rui- pt   = 3110
 3097 - Hiroshi Miura  - ja   = 3097
 3035 - Vitor George   - pt_BR= 3035

Another issue is the quality of the strings. Some source strings are not 
English (but French, Dutch, Polish, etc), have unknown acronyms in it, or


These are bugs and should be fixed.

it's just not clear what the string is about. That mainly affects plugins and 
the imagery. I sometimes open a ticket if the quality is bad. JOSM core 
tickets are resolved promptly, but plugin authors are lazy or don't respond


Yes. That's an issue. The success of JOSM has some drawbacks and 
unmaintained plugins is one of them. The core developers try to manage 
them, but we already have lots of core issues. No solution for this ATM. 
For things like translating issues tickets should nevertheless be opened. 
They are usually easy to fix and thus also easy to handle by the core 
team.


at all. There is no good way to address issues with the imagery strings 
(there is actually no way).


See answers in related ticket #12179.

One solution could be to separate the JOSM core strings from the plugin and 
imagery strings. Then translators can concentrate on the JOSM core, it would 
reduce the timeout errors, and it becomes doable to actually translate all 
(core) strings.


The other solution would be the long discussed migration to transifex. It 
includes also the splitting in 3 string groups.


Plugins are a different topic. I think there is no way to attract people to 
translate a large number of strings that nobody will ever read. An idea would 
be to rank the untranslated strings by the number of plugin installations (do 
we know this?) to start with the important strings. But Launchpad sorts them 
alphabetically, not by importance.


Oh. We know EVERYTHING :-) Don't ask me why this list has reversed order 
to the languages though - probably I implemented them at different 
times...


   4 (* 0.1%) NanoLog
  10 (* 0.2%) matsim
  15 (* 0.2%) videomapping
  22 (* 0.3%) sds
  27 (* 0.4%) globalsat
  32 (* 0.5%) czechaddress
  35 (* 0.6%) michigan_left
  35 (* 0.6%) epci-fr
  37 (* 0.6%) plastic_laf
  38 (* 0.6%) native_password_manager
  41 (* 0.6%) wms-turbo-challenge2
  45 (* 0.7%) canvec_helper
  46 (* 0.7%) scoutsigns
  48 (* 0.8%) OSMRecPlugin
  48 (* 0.8%) surveyor
  50 (* 0.8%) touchscreenhelper
  51 (* 0.8%) o5m
  51 (* 0.8%) pointInfo
  51 (* 0.8%) 

Re: JOSM Translations

2017-02-19 Thread Martin Koppenhoefer
2017-02-19 18:38 GMT+01:00 Holger Mappt :

> It would be interesting to know how important translations are. Maybe it
> is okay for most of the users to use the English version.




if you can understand English it is IMHO preferable to use JOSM in en even
if your language is completely translated, because many terms can be easier
recognized and tags can be found easier if you see them in the actual
language, and you don't face potential translation inconsistencies (e.g. a
node can be a "Knoten", "Node", "Punkt" etc. in German). It also helps to
contribute bug issues, ask a wider audience for specific help, etc. But if
you don't understand English a local translation is obviously needed. One
problem is maybe that translations are not needed by those who are able to
provide them (nobody is scratching his own itch here) ;-)

I admit I have copied the initial message from Dirk to talk, thinking that
indeed there might be people around of those languages with missing
translations, not knowing that help could be needed but willing to give a
hand:
https://lists.openstreetmap.org/pipermail/talk/2017-February/077579.html

I excuse if this was premature...

Cheers,
Martin


Re: JOSM Translations

2017-02-19 Thread simson.gert...@gmail.com

I think the main point is Launchpad. It is completely unusable. Last
time I even needed to request the po file download link twice, just to
check a single translation string. The websearch did't work at all.

The second point is, that we need some groups. I think the users want to
start to translate the presets and the main menu entries and not some
cryptic error messages, they don't understand and therefore are not able
to translate them, even if they want to.

Best Regards,
Stefan

---
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus




Re: JOSM Translations

2017-02-19 Thread Holger Mappt

On 2017-02-19 at 14:19 +0100 Dirk Stöcker wrote:

I had a look at our translation statistics recently and I find it awful,
that there is no real progress to see, compared to the situation a few
years ago.

Does anybody have ideas how more translators could be attracted for the
languages which have many missing texts?

We do daily text uploads and although launchpad has issues from time to
time it usually is stable. The tools exist.

I think the main problem is that people aren't aware they could help and
are needed to do so.


Hi,

It would be interesting to know how important translations are. Maybe it 
is okay for most of the users to use the English version.


I think one of the issues is the enormous number of strings. We reached 
1 a few weeks ago. And there is a pretty high update frequency. You 
need days to catch up if no translation was done for a few months.


Another issue is the quality of the strings. Some source strings are not 
English (but French, Dutch, Polish, etc), have unknown acronyms in it, 
or it's just not clear what the string is about. That mainly affects 
plugins and the imagery. I sometimes open a ticket if the quality is 
bad. JOSM core tickets are resolved promptly, but plugin authors are 
lazy or don't respond at all. There is no good way to address issues 
with the imagery strings (there is actually no way).


Launchpad has issues all the time. It was okay for a while, but now the 
timeout error is back again, I get it every time I translate something. 
What's really missing is the possibility to comment a string to add 
hints for other translators.


One solution could be to separate the JOSM core strings from the plugin 
and imagery strings. Then translators can concentrate on the JOSM core, 
it would reduce the timeout errors, and it becomes doable to actually 
translate all (core) strings.


Plugins are a different topic. I think there is no way to attract people 
to translate a large number of strings that nobody will ever read. An 
idea would be to rank the untranslated strings by the number of plugin 
installations (do we know this?) to start with the important strings. 
But Launchpad sorts them alphabetically, not by importance.


For the imagery strings I opened ticket #12179, but it looks like no one 
is interested to improve the current situation.


Holger





Re: JOSM Translations

2017-02-19 Thread Vincent Privat
First thing is get rid of Launchpad and finish what was started with the
move to Transifex.

The launchpad UI is terrible, especially when you face OOPS errors every
time you search something. I myself completely gave up with the translation
effort because of it, so I presume I am not the only one.

Then one possible outcome would be to promote translation from JOSM itself.
Maybe something like what it is done with Jenkins:
https://wiki.jenkins-ci.org/display/JENKINS/Translation+Assistance+Plugin

2017-02-19 15:31 GMT+01:00 Malcolm Herring :

> On 19/02/2017 13:19, Dirk Stöcker wrote:
>
>> I think the main problem is that people aren't aware they could help and
>> are needed to do so.
>>
>
> Maybe post your message in the general list? There may be able translators
> there who do not read this newsgroup.
>
>
>


Re: JOSM Translations

2017-02-19 Thread Malcolm Herring

On 19/02/2017 13:19, Dirk Stöcker wrote:

I think the main problem is that people aren't aware they could help and
are needed to do so.


Maybe post your message in the general list? There may be able 
translators there who do not read this newsgroup.